README.md

---
language:
- en
license: mit
license_link: https://huggingface.co/datasets/hcltech-robotics/simreadybench/blob/main/NVIDIA-ASSET-LICENSE.txt
pretty_name: SimReadyBench
thumbnail: https://huggingface.co/datasets/hcltech-robotics/simreadybench/resolve/main/media/social-card.png
authors:
- name: Chris von Csefalvay
  affiliation: HCLTech
- name: Pranav Doma
  affiliation: HCLTech
- name: Achyuthan Unni Krishnan
  affiliation: HCLTech
- name: Tamas Foldi
  affiliation: HCLTech
configs:
- config_name: all
  default: true
  data_files:
  - split: test
    path:
    - viewer/nv25-gpt-6.1-sol-high-pi-none-turn64-20261020/scores.parquet
    - viewer/nv25-qwen38-27b-medium-pi-none-turn64-20261021/scores.parquet
    - viewer/nv25-qwen38-27b-xhigh-pi-none-turn64-20261021/scores.parquet
  features:
  - name: reference_image
    dtype: image
  - name: candidate_image
    dtype: image
  - name: capture_view
    dtype: string
  - name: condition_id
    dtype: string
  - name: asset_id
    dtype: string
  - name: requested_model
    dtype: string
  - name: harness_id
    dtype: string
  - name: skill_set_id
    dtype: string
  - name: repetition
    dtype: int64
  - name: status
    dtype: string
  - name: formal_fidelity_percent
    dtype: float64
  - name: visual_fidelity_percent
    dtype: float64
  - name: kinematic_fidelity_percent
    dtype: float64
  - name: simready_reward
    dtype: float64
  - name: mandatory_passed
    dtype: float64
  - name: mandatory_total
    dtype: float64
  - name: mandatory_failed
    dtype: float64
  - name: mandatory_unknown
    dtype: float64
  - name: mandatory_lower_bound_percent
    dtype: float64
  - name: mandatory_upper_bound_percent
    dtype: float64
  - name: completion_bonus
    dtype: float64
  - name: optional_score_percent
    dtype: float64
  - name: optional_passed
    dtype: float64
  - name: optional_failed
    dtype: float64
  - name: optional_unknown
    dtype: float64
  - name: optional_total
    dtype: float64
  - name: optional_measured
    dtype: float64
  - name: optional_coverage
    dtype: float64
  - name: optional_lower_bound_percent
    dtype: float64
  - name: optional_upper_bound_percent
    dtype: float64
  - name: fidelity_shape_percent
    dtype: float64
  - name: fidelity_rigid_parts_percent
    dtype: float64
  - name: fidelity_joints_percent
    dtype: float64
  - name: fidelity_materials_percent
    dtype: float64
  - name: fidelity_shape_coverage
    dtype: float64
  - name: fidelity_rigid_parts_coverage
    dtype: float64
  - name: fidelity_joints_coverage
    dtype: float64
  - name: fidelity_materials_coverage
    dtype: float64
  - name: fidelity_reference_geometry_status
    dtype: string
  - name: formal_fidelity_status
    dtype: string
  - name: formal_fidelity_earned
    dtype: float64
  - name: formal_fidelity_eligible
    dtype: float64
  - name: formal_fidelity_measured
    dtype: float64
  - name: formal_fidelity_coverage
    dtype: float64
  - name: formal_fidelity_lower_bound_percent
    dtype: float64
  - name: formal_fidelity_upper_bound_percent
    dtype: float64
  - name: visual_fidelity_status
    dtype: string
  - name: visual_fidelity_earned
    dtype: float64
  - name: visual_fidelity_eligible
    dtype: float64
  - name: visual_fidelity_measured
    dtype: float64
  - name: visual_fidelity_coverage
    dtype: float64
  - name: visual_fidelity_lower_bound_percent
    dtype: float64
  - name: visual_fidelity_upper_bound_percent
    dtype: float64
  - name: kinematic_fidelity_status
    dtype: string
  - name: kinematic_fidelity_earned
    dtype: float64
  - name: kinematic_fidelity_eligible
    dtype: float64
  - name: kinematic_fidelity_measured
    dtype: float64
  - name: kinematic_fidelity_coverage
    dtype: float64
  - name: kinematic_fidelity_lower_bound_percent
    dtype: float64
  - name: kinematic_fidelity_upper_bound_percent
    dtype: float64
  - name: visual_appearance_percent
    dtype: float64
  - name: visual_appearance_status
    dtype: string
  - name: visual_appearance_earned
    dtype: float64
  - name: visual_appearance_eligible
    dtype: float64
  - name: visual_appearance_measured
    dtype: float64
  - name: visual_appearance_coverage
    dtype: float64
  - name: visual_appearance_lower_bound_percent
    dtype: float64
  - name: visual_appearance_upper_bound_percent
    dtype: float64
  - name: kinematic_reference_joint_count
    dtype: float64
  - name: kinematic_candidate_joint_count
    dtype: float64
  - name: kinematic_matched_joint_count
    dtype: float64
  - name: kinematic_missing_joint_count
    dtype: float64
  - name: kinematic_extra_joint_count
    dtype: float64
  - name: kinematic_moving_joints_percent
    dtype: float64
  - name: kinematic_moving_joints_status
    dtype: string
  - name: kinematic_moving_joints_earned
    dtype: float64
  - name: kinematic_moving_joints_eligible
    dtype: float64
  - name: kinematic_moving_joints_measured
    dtype: float64
  - name: kinematic_moving_joints_coverage
    dtype: float64
  - name: kinematic_moving_joints_lower_bound_percent
    dtype: float64
  - name: kinematic_moving_joints_upper_bound_percent
    dtype: float64
  - name: kinematic_fixed_constraints_percent
    dtype: float64
  - name: kinematic_fixed_constraints_status
    dtype: string
  - name: kinematic_fixed_constraints_earned
    dtype: float64
  - name: kinematic_fixed_constraints_eligible
    dtype: float64
  - name: kinematic_fixed_constraints_measured
    dtype: float64
  - name: kinematic_fixed_constraints_coverage
    dtype: float64
  - name: kinematic_fixed_constraints_lower_bound_percent
    dtype: float64
  - name: kinematic_fixed_constraints_upper_bound_percent
    dtype: float64
  - name: fidelity_shape_earned
    dtype: float64
  - name: fidelity_shape_eligible
    dtype: float64
  - name: fidelity_shape_measured
    dtype: float64
  - name: fidelity_shape_status
    dtype: string
  - name: fidelity_rigid_parts_earned
    dtype: float64
  - name: fidelity_rigid_parts_eligible
    dtype: float64
  - name: fidelity_rigid_parts_measured
    dtype: float64
  - name: fidelity_rigid_parts_status
    dtype: string
  - name: fidelity_joints_earned
    dtype: float64
  - name: fidelity_joints_eligible
    dtype: float64
  - name: fidelity_joints_measured
    dtype: float64
  - name: fidelity_joints_status
    dtype: string
  - name: fidelity_materials_earned
    dtype: float64
  - name: fidelity_materials_eligible
    dtype: float64
  - name: fidelity_materials_measured
    dtype: float64
  - name: fidelity_materials_status
    dtype: string
  - name: fidelity_shape_precision
    dtype: float64
  - name: fidelity_shape_recall
    dtype: float64
  - name: fidelity_shape_samples_each_direction
    dtype: float64
  - name: fidelity_rigid_parts_geometry_percent
    dtype: float64
  - name: fidelity_rigid_parts_structure_percent
    dtype: float64
  - name: fidelity_rigid_parts_structure_earned
    dtype: float64
  - name: fidelity_rigid_parts_structure_eligible
    dtype: float64
  - name: fidelity_joints_reference_count
    dtype: float64
  - name: fidelity_joints_candidate_count
    dtype: float64
  - name: inference_calls_used
    dtype: float64
  - name: validation_calls_used
    dtype: float64
  - name: model_request_errors
    dtype: float64
  - name: elapsed_seconds
    dtype: float64
  - name: model_request_seconds
    dtype: float64
  - name: input_tokens
    dtype: float64
  - name: output_tokens
    dtype: float64
  - name: total_tokens
    dtype: float64
  - name: inference_cost_usd
    dtype: float64
  - name: compute_cost_usd
    dtype: float64
  - name: total_cost_usd
    dtype: float64
  - name: authoring_protocol
    dtype: string
  - name: max_output_tokens
    dtype: int64
  - name: output_limit
    dtype: string
  - name: reasoning_effort
    dtype: string
  - name: run_id
    dtype: string
- config_name: by_model
  features:
  - name: reference_image
    dtype: image
  - name: candidate_image
    dtype: image
  - name: capture_view
    dtype: string
  - name: condition_id
    dtype: string
  - name: asset_id
    dtype: string
  - name: requested_model
    dtype: string
  - name: harness_id
    dtype: string
  - name: skill_set_id
    dtype: string
  - name: repetition
    dtype: int64
  - name: status
    dtype: string
  - name: formal_fidelity_percent
    dtype: float64
  - name: visual_fidelity_percent
    dtype: float64
  - name: kinematic_fidelity_percent
    dtype: float64
  - name: simready_reward
    dtype: float64
  - name: mandatory_passed
    dtype: float64
  - name: mandatory_total
    dtype: float64
  - name: mandatory_failed
    dtype: float64
  - name: mandatory_unknown
    dtype: float64
  - name: mandatory_lower_bound_percent
    dtype: float64
  - name: mandatory_upper_bound_percent
    dtype: float64
  - name: completion_bonus
    dtype: float64
  - name: optional_score_percent
    dtype: float64
  - name: optional_passed
    dtype: float64
  - name: optional_failed
    dtype: float64
  - name: optional_unknown
    dtype: float64
  - name: optional_total
    dtype: float64
  - name: optional_measured
    dtype: float64
  - name: optional_coverage
    dtype: float64
  - name: optional_lower_bound_percent
    dtype: float64
  - name: optional_upper_bound_percent
    dtype: float64
  - name: fidelity_shape_percent
    dtype: float64
  - name: fidelity_rigid_parts_percent
    dtype: float64
  - name: fidelity_joints_percent
    dtype: float64
  - name: fidelity_materials_percent
    dtype: float64
  - name: fidelity_shape_coverage
    dtype: float64
  - name: fidelity_rigid_parts_coverage
    dtype: float64
  - name: fidelity_joints_coverage
    dtype: float64
  - name: fidelity_materials_coverage
    dtype: float64
  - name: fidelity_reference_geometry_status
    dtype: string
  - name: formal_fidelity_status
    dtype: string
  - name: formal_fidelity_earned
    dtype: float64
  - name: formal_fidelity_eligible
    dtype: float64
  - name: formal_fidelity_measured
    dtype: float64
  - name: formal_fidelity_coverage
    dtype: float64
  - name: formal_fidelity_lower_bound_percent
    dtype: float64
  - name: formal_fidelity_upper_bound_percent
    dtype: float64
  - name: visual_fidelity_status
    dtype: string
  - name: visual_fidelity_earned
    dtype: float64
  - name: visual_fidelity_eligible
    dtype: float64
  - name: visual_fidelity_measured
    dtype: float64
  - name: visual_fidelity_coverage
    dtype: float64
  - name: visual_fidelity_lower_bound_percent
    dtype: float64
  - name: visual_fidelity_upper_bound_percent
    dtype: float64
  - name: kinematic_fidelity_status
    dtype: string
  - name: kinematic_fidelity_earned
    dtype: float64
  - name: kinematic_fidelity_eligible
    dtype: float64
  - name: kinematic_fidelity_measured
    dtype: float64
  - name: kinematic_fidelity_coverage
    dtype: float64
  - name: kinematic_fidelity_lower_bound_percent
    dtype: float64
  - name: kinematic_fidelity_upper_bound_percent
    dtype: float64
  - name: visual_appearance_percent
    dtype: float64
  - name: visual_appearance_status
    dtype: string
  - name: visual_appearance_earned
    dtype: float64
  - name: visual_appearance_eligible
    dtype: float64
  - name: visual_appearance_measured
    dtype: float64
  - name: visual_appearance_coverage
    dtype: float64
  - name: visual_appearance_lower_bound_percent
    dtype: float64
  - name: visual_appearance_upper_bound_percent
    dtype: float64
  - name: kinematic_reference_joint_count
    dtype: float64
  - name: kinematic_candidate_joint_count
    dtype: float64
  - name: kinematic_matched_joint_count
    dtype: float64
  - name: kinematic_missing_joint_count
    dtype: float64
  - name: kinematic_extra_joint_count
    dtype: float64
  - name: kinematic_moving_joints_percent
    dtype: float64
  - name: kinematic_moving_joints_status
    dtype: string
  - name: kinematic_moving_joints_earned
    dtype: float64
  - name: kinematic_moving_joints_eligible
    dtype: float64
  - name: kinematic_moving_joints_measured
    dtype: float64
  - name: kinematic_moving_joints_coverage
    dtype: float64
  - name: kinematic_moving_joints_lower_bound_percent
    dtype: float64
  - name: kinematic_moving_joints_upper_bound_percent
    dtype: float64
  - name: kinematic_fixed_constraints_percent
    dtype: float64
  - name: kinematic_fixed_constraints_status
    dtype: string
  - name: kinematic_fixed_constraints_earned
    dtype: float64
  - name: kinematic_fixed_constraints_eligible
    dtype: float64
  - name: kinematic_fixed_constraints_measured
    dtype: float64
  - name: kinematic_fixed_constraints_coverage
    dtype: float64
  - name: kinematic_fixed_constraints_lower_bound_percent
    dtype: float64
  - name: kinematic_fixed_constraints_upper_bound_percent
    dtype: float64
  - name: fidelity_shape_earned
    dtype: float64
  - name: fidelity_shape_eligible
    dtype: float64
  - name: fidelity_shape_measured
    dtype: float64
  - name: fidelity_shape_status
    dtype: string
  - name: fidelity_rigid_parts_earned
    dtype: float64
  - name: fidelity_rigid_parts_eligible
    dtype: float64
  - name: fidelity_rigid_parts_measured
    dtype: float64
  - name: fidelity_rigid_parts_status
    dtype: string
  - name: fidelity_joints_earned
    dtype: float64
  - name: fidelity_joints_eligible
    dtype: float64
  - name: fidelity_joints_measured
    dtype: float64
  - name: fidelity_joints_status
    dtype: string
  - name: fidelity_materials_earned
    dtype: float64
  - name: fidelity_materials_eligible
    dtype: float64
  - name: fidelity_materials_measured
    dtype: float64
  - name: fidelity_materials_status
    dtype: string
  - name: fidelity_shape_precision
    dtype: float64
  - name: fidelity_shape_recall
    dtype: float64
  - name: fidelity_shape_samples_each_direction
    dtype: float64
  - name: fidelity_rigid_parts_geometry_percent
    dtype: float64
  - name: fidelity_rigid_parts_structure_percent
    dtype: float64
  - name: fidelity_rigid_parts_structure_earned
    dtype: float64
  - name: fidelity_rigid_parts_structure_eligible
    dtype: float64
  - name: fidelity_joints_reference_count
    dtype: float64
  - name: fidelity_joints_candidate_count
    dtype: float64
  - name: inference_calls_used
    dtype: float64
  - name: validation_calls_used
    dtype: float64
  - name: model_request_errors
    dtype: float64
  - name: elapsed_seconds
    dtype: float64
  - name: model_request_seconds
    dtype: float64
  - name: input_tokens
    dtype: float64
  - name: output_tokens
    dtype: float64
  - name: total_tokens
    dtype: float64
  - name: inference_cost_usd
    dtype: float64
  - name: compute_cost_usd
    dtype: float64
  - name: total_cost_usd
    dtype: float64
  - name: authoring_protocol
    dtype: string
  - name: max_output_tokens
    dtype: int64
  - name: output_limit
    dtype: string
  - name: reasoning_effort
    dtype: string
  - name: run_id
    dtype: string
  data_files:
  - split: gpt_6_1_sol
    path:
    - viewer/nv25-gpt-6.1-sol-high-pi-none-turn64-20261020/scores.parquet
  - split: qwen_qwen3_8_27b_deepinfra
    path:
    - viewer/nv25-qwen38-27b-medium-pi-none-turn64-20261021/scores.parquet
    - viewer/nv25-qwen38-27b-xhigh-pi-none-turn64-20261021/scores.parquet
- config_name: by_harness
  features:
  - name: reference_image
    dtype: image
  - name: candidate_image
    dtype: image
  - name: capture_view
    dtype: string
  - name: condition_id
    dtype: string
  - name: asset_id
    dtype: string
  - name: requested_model
    dtype: string
  - name: harness_id
    dtype: string
  - name: skill_set_id
    dtype: string
  - name: repetition
    dtype: int64
  - name: status
    dtype: string
  - name: formal_fidelity_percent
    dtype: float64
  - name: visual_fidelity_percent
    dtype: float64
  - name: kinematic_fidelity_percent
    dtype: float64
  - name: simready_reward
    dtype: float64
  - name: mandatory_passed
    dtype: float64
  - name: mandatory_total
    dtype: float64
  - name: mandatory_failed
    dtype: float64
  - name: mandatory_unknown
    dtype: float64
  - name: mandatory_lower_bound_percent
    dtype: float64
  - name: mandatory_upper_bound_percent
    dtype: float64
  - name: completion_bonus
    dtype: float64
  - name: optional_score_percent
    dtype: float64
  - name: optional_passed
    dtype: float64
  - name: optional_failed
    dtype: float64
  - name: optional_unknown
    dtype: float64
  - name: optional_total
    dtype: float64
  - name: optional_measured
    dtype: float64
  - name: optional_coverage
    dtype: float64
  - name: optional_lower_bound_percent
    dtype: float64
  - name: optional_upper_bound_percent
    dtype: float64
  - name: fidelity_shape_percent
    dtype: float64
  - name: fidelity_rigid_parts_percent
    dtype: float64
  - name: fidelity_joints_percent
    dtype: float64
  - name: fidelity_materials_percent
    dtype: float64
  - name: fidelity_shape_coverage
    dtype: float64
  - name: fidelity_rigid_parts_coverage
    dtype: float64
  - name: fidelity_joints_coverage
    dtype: float64
  - name: fidelity_materials_coverage
    dtype: float64
  - name: fidelity_reference_geometry_status
    dtype: string
  - name: formal_fidelity_status
    dtype: string
  - name: formal_fidelity_earned
    dtype: float64
  - name: formal_fidelity_eligible
    dtype: float64
  - name: formal_fidelity_measured
    dtype: float64
  - name: formal_fidelity_coverage
    dtype: float64
  - name: formal_fidelity_lower_bound_percent
    dtype: float64
  - name: formal_fidelity_upper_bound_percent
    dtype: float64
  - name: visual_fidelity_status
    dtype: string
  - name: visual_fidelity_earned
    dtype: float64
  - name: visual_fidelity_eligible
    dtype: float64
  - name: visual_fidelity_measured
    dtype: float64
  - name: visual_fidelity_coverage
    dtype: float64
  - name: visual_fidelity_lower_bound_percent
    dtype: float64
  - name: visual_fidelity_upper_bound_percent
    dtype: float64
  - name: kinematic_fidelity_status
    dtype: string
  - name: kinematic_fidelity_earned
    dtype: float64
  - name: kinematic_fidelity_eligible
    dtype: float64
  - name: kinematic_fidelity_measured
    dtype: float64
  - name: kinematic_fidelity_coverage
    dtype: float64
  - name: kinematic_fidelity_lower_bound_percent
    dtype: float64
  - name: kinematic_fidelity_upper_bound_percent
    dtype: float64
  - name: visual_appearance_percent
    dtype: float64
  - name: visual_appearance_status
    dtype: string
  - name: visual_appearance_earned
    dtype: float64
  - name: visual_appearance_eligible
    dtype: float64
  - name: visual_appearance_measured
    dtype: float64
  - name: visual_appearance_coverage
    dtype: float64
  - name: visual_appearance_lower_bound_percent
    dtype: float64
  - name: visual_appearance_upper_bound_percent
    dtype: float64
  - name: kinematic_reference_joint_count
    dtype: float64
  - name: kinematic_candidate_joint_count
    dtype: float64
  - name: kinematic_matched_joint_count
    dtype: float64
  - name: kinematic_missing_joint_count
    dtype: float64
  - name: kinematic_extra_joint_count
    dtype: float64
  - name: kinematic_moving_joints_percent
    dtype: float64
  - name: kinematic_moving_joints_status
    dtype: string
  - name: kinematic_moving_joints_earned
    dtype: float64
  - name: kinematic_moving_joints_eligible
    dtype: float64
  - name: kinematic_moving_joints_measured
    dtype: float64
  - name: kinematic_moving_joints_coverage
    dtype: float64
  - name: kinematic_moving_joints_lower_bound_percent
    dtype: float64
  - name: kinematic_moving_joints_upper_bound_percent
    dtype: float64
  - name: kinematic_fixed_constraints_percent
    dtype: float64
  - name: kinematic_fixed_constraints_status
    dtype: string
  - name: kinematic_fixed_constraints_earned
    dtype: float64
  - name: kinematic_fixed_constraints_eligible
    dtype: float64
  - name: kinematic_fixed_constraints_measured
    dtype: float64
  - name: kinematic_fixed_constraints_coverage
    dtype: float64
  - name: kinematic_fixed_constraints_lower_bound_percent
    dtype: float64
  - name: kinematic_fixed_constraints_upper_bound_percent
    dtype: float64
  - name: fidelity_shape_earned
    dtype: float64
  - name: fidelity_shape_eligible
    dtype: float64
  - name: fidelity_shape_measured
    dtype: float64
  - name: fidelity_shape_status
    dtype: string
  - name: fidelity_rigid_parts_earned
    dtype: float64
  - name: fidelity_rigid_parts_eligible
    dtype: float64
  - name: fidelity_rigid_parts_measured
    dtype: float64
  - name: fidelity_rigid_parts_status
    dtype: string
  - name: fidelity_joints_earned
    dtype: float64
  - name: fidelity_joints_eligible
    dtype: float64
  - name: fidelity_joints_measured
    dtype: float64
  - name: fidelity_joints_status
    dtype: string
  - name: fidelity_materials_earned
    dtype: float64
  - name: fidelity_materials_eligible
    dtype: float64
  - name: fidelity_materials_measured
    dtype: float64
  - name: fidelity_materials_status
    dtype: string
  - name: fidelity_shape_precision
    dtype: float64
  - name: fidelity_shape_recall
    dtype: float64
  - name: fidelity_shape_samples_each_direction
    dtype: float64
  - name: fidelity_rigid_parts_geometry_percent
    dtype: float64
  - name: fidelity_rigid_parts_structure_percent
    dtype: float64
  - name: fidelity_rigid_parts_structure_earned
    dtype: float64
  - name: fidelity_rigid_parts_structure_eligible
    dtype: float64
  - name: fidelity_joints_reference_count
    dtype: float64
  - name: fidelity_joints_candidate_count
    dtype: float64
  - name: inference_calls_used
    dtype: float64
  - name: validation_calls_used
    dtype: float64
  - name: model_request_errors
    dtype: float64
  - name: elapsed_seconds
    dtype: float64
  - name: model_request_seconds
    dtype: float64
  - name: input_tokens
    dtype: float64
  - name: output_tokens
    dtype: float64
  - name: total_tokens
    dtype: float64
  - name: inference_cost_usd
    dtype: float64
  - name: compute_cost_usd
    dtype: float64
  - name: total_cost_usd
    dtype: float64
  - name: authoring_protocol
    dtype: string
  - name: max_output_tokens
    dtype: int64
  - name: output_limit
    dtype: string
  - name: reasoning_effort
    dtype: string
  - name: run_id
    dtype: string
  data_files:
  - split: pi
    path:
    - viewer/nv25-gpt-6.1-sol-high-pi-none-turn64-20261020/scores.parquet
    - viewer/nv25-qwen38-27b-medium-pi-none-turn64-20261021/scores.parquet
    - viewer/nv25-qwen38-27b-xhigh-pi-none-turn64-20261021/scores.parquet
- config_name: by_skills
  features:
  - name: reference_image
    dtype: image
  - name: candidate_image
    dtype: image
  - name: capture_view
    dtype: string
  - name: condition_id
    dtype: string
  - name: asset_id
    dtype: string
  - name: requested_model
    dtype: string
  - name: harness_id
    dtype: string
  - name: skill_set_id
    dtype: string
  - name: repetition
    dtype: int64
  - name: status
    dtype: string
  - name: formal_fidelity_percent
    dtype: float64
  - name: visual_fidelity_percent
    dtype: float64
  - name: kinematic_fidelity_percent
    dtype: float64
  - name: simready_reward
    dtype: float64
  - name: mandatory_passed
    dtype: float64
  - name: mandatory_total
    dtype: float64
  - name: mandatory_failed
    dtype: float64
  - name: mandatory_unknown
    dtype: float64
  - name: mandatory_lower_bound_percent
    dtype: float64
  - name: mandatory_upper_bound_percent
    dtype: float64
  - name: completion_bonus
    dtype: float64
  - name: optional_score_percent
    dtype: float64
  - name: optional_passed
    dtype: float64
  - name: optional_failed
    dtype: float64
  - name: optional_unknown
    dtype: float64
  - name: optional_total
    dtype: float64
  - name: optional_measured
    dtype: float64
  - name: optional_coverage
    dtype: float64
  - name: optional_lower_bound_percent
    dtype: float64
  - name: optional_upper_bound_percent
    dtype: float64
  - name: fidelity_shape_percent
    dtype: float64
  - name: fidelity_rigid_parts_percent
    dtype: float64
  - name: fidelity_joints_percent
    dtype: float64
  - name: fidelity_materials_percent
    dtype: float64
  - name: fidelity_shape_coverage
    dtype: float64
  - name: fidelity_rigid_parts_coverage
    dtype: float64
  - name: fidelity_joints_coverage
    dtype: float64
  - name: fidelity_materials_coverage
    dtype: float64
  - name: fidelity_reference_geometry_status
    dtype: string
  - name: formal_fidelity_status
    dtype: string
  - name: formal_fidelity_earned
    dtype: float64
  - name: formal_fidelity_eligible
    dtype: float64
  - name: formal_fidelity_measured
    dtype: float64
  - name: formal_fidelity_coverage
    dtype: float64
  - name: formal_fidelity_lower_bound_percent
    dtype: float64
  - name: formal_fidelity_upper_bound_percent
    dtype: float64
  - name: visual_fidelity_status
    dtype: string
  - name: visual_fidelity_earned
    dtype: float64
  - name: visual_fidelity_eligible
    dtype: float64
  - name: visual_fidelity_measured
    dtype: float64
  - name: visual_fidelity_coverage
    dtype: float64
  - name: visual_fidelity_lower_bound_percent
    dtype: float64
  - name: visual_fidelity_upper_bound_percent
    dtype: float64
  - name: kinematic_fidelity_status
    dtype: string
  - name: kinematic_fidelity_earned
    dtype: float64
  - name: kinematic_fidelity_eligible
    dtype: float64
  - name: kinematic_fidelity_measured
    dtype: float64
  - name: kinematic_fidelity_coverage
    dtype: float64
  - name: kinematic_fidelity_lower_bound_percent
    dtype: float64
  - name: kinematic_fidelity_upper_bound_percent
    dtype: float64
  - name: visual_appearance_percent
    dtype: float64
  - name: visual_appearance_status
    dtype: string
  - name: visual_appearance_earned
    dtype: float64
  - name: visual_appearance_eligible
    dtype: float64
  - name: visual_appearance_measured
    dtype: float64
  - name: visual_appearance_coverage
    dtype: float64
  - name: visual_appearance_lower_bound_percent
    dtype: float64
  - name: visual_appearance_upper_bound_percent
    dtype: float64
  - name: kinematic_reference_joint_count
    dtype: float64
  - name: kinematic_candidate_joint_count
    dtype: float64
  - name: kinematic_matched_joint_count
    dtype: float64
  - name: kinematic_missing_joint_count
    dtype: float64
  - name: kinematic_extra_joint_count
    dtype: float64
  - name: kinematic_moving_joints_percent
    dtype: float64
  - name: kinematic_moving_joints_status
    dtype: string
  - name: kinematic_moving_joints_earned
    dtype: float64
  - name: kinematic_moving_joints_eligible
    dtype: float64
  - name: kinematic_moving_joints_measured
    dtype: float64
  - name: kinematic_moving_joints_coverage
    dtype: float64
  - name: kinematic_moving_joints_lower_bound_percent
    dtype: float64
  - name: kinematic_moving_joints_upper_bound_percent
    dtype: float64
  - name: kinematic_fixed_constraints_percent
    dtype: float64
  - name: kinematic_fixed_constraints_status
    dtype: string
  - name: kinematic_fixed_constraints_earned
    dtype: float64
  - name: kinematic_fixed_constraints_eligible
    dtype: float64
  - name: kinematic_fixed_constraints_measured
    dtype: float64
  - name: kinematic_fixed_constraints_coverage
    dtype: float64
  - name: kinematic_fixed_constraints_lower_bound_percent
    dtype: float64
  - name: kinematic_fixed_constraints_upper_bound_percent
    dtype: float64
  - name: fidelity_shape_earned
    dtype: float64
  - name: fidelity_shape_eligible
    dtype: float64
  - name: fidelity_shape_measured
    dtype: float64
  - name: fidelity_shape_status
    dtype: string
  - name: fidelity_rigid_parts_earned
    dtype: float64
  - name: fidelity_rigid_parts_eligible
    dtype: float64
  - name: fidelity_rigid_parts_measured
    dtype: float64
  - name: fidelity_rigid_parts_status
    dtype: string
  - name: fidelity_joints_earned
    dtype: float64
  - name: fidelity_joints_eligible
    dtype: float64
  - name: fidelity_joints_measured
    dtype: float64
  - name: fidelity_joints_status
    dtype: string
  - name: fidelity_materials_earned
    dtype: float64
  - name: fidelity_materials_eligible
    dtype: float64
  - name: fidelity_materials_measured
    dtype: float64
  - name: fidelity_materials_status
    dtype: string
  - name: fidelity_shape_precision
    dtype: float64
  - name: fidelity_shape_recall
    dtype: float64
  - name: fidelity_shape_samples_each_direction
    dtype: float64
  - name: fidelity_rigid_parts_geometry_percent
    dtype: float64
  - name: fidelity_rigid_parts_structure_percent
    dtype: float64
  - name: fidelity_rigid_parts_structure_earned
    dtype: float64
  - name: fidelity_rigid_parts_structure_eligible
    dtype: float64
  - name: fidelity_joints_reference_count
    dtype: float64
  - name: fidelity_joints_candidate_count
    dtype: float64
  - name: inference_calls_used
    dtype: float64
  - name: validation_calls_used
    dtype: float64
  - name: model_request_errors
    dtype: float64
  - name: elapsed_seconds
    dtype: float64
  - name: model_request_seconds
    dtype: float64
  - name: input_tokens
    dtype: float64
  - name: output_tokens
    dtype: float64
  - name: total_tokens
    dtype: float64
  - name: inference_cost_usd
    dtype: float64
  - name: compute_cost_usd
    dtype: float64
  - name: total_cost_usd
    dtype: float64
  - name: authoring_protocol
    dtype: string
  - name: max_output_tokens
    dtype: int64
  - name: output_limit
    dtype: string
  - name: reasoning_effort
    dtype: string
  - name: run_id
    dtype: string
  data_files:
  - split: none
    path:
    - viewer/nv25-gpt-6.1-sol-high-pi-none-turn64-20261020/scores.parquet
    - viewer/nv25-qwen38-27b-medium-pi-none-turn64-20261021/scores.parquet
    - viewer/nv25-qwen38-27b-xhigh-pi-none-turn64-20261021/scores.parquet
- config_name: by_reference_set
  features:
  - name: reference_image
    dtype: image
  - name: candidate_image
    dtype: image
  - name: capture_view
    dtype: string
  - name: condition_id
    dtype: string
  - name: asset_id
    dtype: string
  - name: requested_model
    dtype: string
  - name: harness_id
    dtype: string
  - name: skill_set_id
    dtype: string
  - name: repetition
    dtype: int64
  - name: status
    dtype: string
  - name: formal_fidelity_percent
    dtype: float64
  - name: visual_fidelity_percent
    dtype: float64
  - name: kinematic_fidelity_percent
    dtype: float64
  - name: simready_reward
    dtype: float64
  - name: mandatory_passed
    dtype: float64
  - name: mandatory_total
    dtype: float64
  - name: mandatory_failed
    dtype: float64
  - name: mandatory_unknown
    dtype: float64
  - name: mandatory_lower_bound_percent
    dtype: float64
  - name: mandatory_upper_bound_percent
    dtype: float64
  - name: completion_bonus
    dtype: float64
  - name: optional_score_percent
    dtype: float64
  - name: optional_passed
    dtype: float64
  - name: optional_failed
    dtype: float64
  - name: optional_unknown
    dtype: float64
  - name: optional_total
    dtype: float64
  - name: optional_measured
    dtype: float64
  - name: optional_coverage
    dtype: float64
  - name: optional_lower_bound_percent
    dtype: float64
  - name: optional_upper_bound_percent
    dtype: float64
  - name: fidelity_shape_percent
    dtype: float64
  - name: fidelity_rigid_parts_percent
    dtype: float64
  - name: fidelity_joints_percent
    dtype: float64
  - name: fidelity_materials_percent
    dtype: float64
  - name: fidelity_shape_coverage
    dtype: float64
  - name: fidelity_rigid_parts_coverage
    dtype: float64
  - name: fidelity_joints_coverage
    dtype: float64
  - name: fidelity_materials_coverage
    dtype: float64
  - name: fidelity_reference_geometry_status
    dtype: string
  - name: formal_fidelity_status
    dtype: string
  - name: formal_fidelity_earned
    dtype: float64
  - name: formal_fidelity_eligible
    dtype: float64
  - name: formal_fidelity_measured
    dtype: float64
  - name: formal_fidelity_coverage
    dtype: float64
  - name: formal_fidelity_lower_bound_percent
    dtype: float64
  - name: formal_fidelity_upper_bound_percent
    dtype: float64
  - name: visual_fidelity_status
    dtype: string
  - name: visual_fidelity_earned
    dtype: float64
  - name: visual_fidelity_eligible
    dtype: float64
  - name: visual_fidelity_measured
    dtype: float64
  - name: visual_fidelity_coverage
    dtype: float64
  - name: visual_fidelity_lower_bound_percent
    dtype: float64
  - name: visual_fidelity_upper_bound_percent
    dtype: float64
  - name: kinematic_fidelity_status
    dtype: string
  - name: kinematic_fidelity_earned
    dtype: float64
  - name: kinematic_fidelity_eligible
    dtype: float64
  - name: kinematic_fidelity_measured
    dtype: float64
  - name: kinematic_fidelity_coverage
    dtype: float64
  - name: kinematic_fidelity_lower_bound_percent
    dtype: float64
  - name: kinematic_fidelity_upper_bound_percent
    dtype: float64
  - name: visual_appearance_percent
    dtype: float64
  - name: visual_appearance_status
    dtype: string
  - name: visual_appearance_earned
    dtype: float64
  - name: visual_appearance_eligible
    dtype: float64
  - name: visual_appearance_measured
    dtype: float64
  - name: visual_appearance_coverage
    dtype: float64
  - name: visual_appearance_lower_bound_percent
    dtype: float64
  - name: visual_appearance_upper_bound_percent
    dtype: float64
  - name: kinematic_reference_joint_count
    dtype: float64
  - name: kinematic_candidate_joint_count
    dtype: float64
  - name: kinematic_matched_joint_count
    dtype: float64
  - name: kinematic_missing_joint_count
    dtype: float64
  - name: kinematic_extra_joint_count
    dtype: float64
  - name: kinematic_moving_joints_percent
    dtype: float64
  - name: kinematic_moving_joints_status
    dtype: string
  - name: kinematic_moving_joints_earned
    dtype: float64
  - name: kinematic_moving_joints_eligible
    dtype: float64
  - name: kinematic_moving_joints_measured
    dtype: float64
  - name: kinematic_moving_joints_coverage
    dtype: float64
  - name: kinematic_moving_joints_lower_bound_percent
    dtype: float64
  - name: kinematic_moving_joints_upper_bound_percent
    dtype: float64
  - name: kinematic_fixed_constraints_percent
    dtype: float64
  - name: kinematic_fixed_constraints_status
    dtype: string
  - name: kinematic_fixed_constraints_earned
    dtype: float64
  - name: kinematic_fixed_constraints_eligible
    dtype: float64
  - name: kinematic_fixed_constraints_measured
    dtype: float64
  - name: kinematic_fixed_constraints_coverage
    dtype: float64
  - name: kinematic_fixed_constraints_lower_bound_percent
    dtype: float64
  - name: kinematic_fixed_constraints_upper_bound_percent
    dtype: float64
  - name: fidelity_shape_earned
    dtype: float64
  - name: fidelity_shape_eligible
    dtype: float64
  - name: fidelity_shape_measured
    dtype: float64
  - name: fidelity_shape_status
    dtype: string
  - name: fidelity_rigid_parts_earned
    dtype: float64
  - name: fidelity_rigid_parts_eligible
    dtype: float64
  - name: fidelity_rigid_parts_measured
    dtype: float64
  - name: fidelity_rigid_parts_status
    dtype: string
  - name: fidelity_joints_earned
    dtype: float64
  - name: fidelity_joints_eligible
    dtype: float64
  - name: fidelity_joints_measured
    dtype: float64
  - name: fidelity_joints_status
    dtype: string
  - name: fidelity_materials_earned
    dtype: float64
  - name: fidelity_materials_eligible
    dtype: float64
  - name: fidelity_materials_measured
    dtype: float64
  - name: fidelity_materials_status
    dtype: string
  - name: fidelity_shape_precision
    dtype: float64
  - name: fidelity_shape_recall
    dtype: float64
  - name: fidelity_shape_samples_each_direction
    dtype: float64
  - name: fidelity_rigid_parts_geometry_percent
    dtype: float64
  - name: fidelity_rigid_parts_structure_percent
    dtype: float64
  - name: fidelity_rigid_parts_structure_earned
    dtype: float64
  - name: fidelity_rigid_parts_structure_eligible
    dtype: float64
  - name: fidelity_joints_reference_count
    dtype: float64
  - name: fidelity_joints_candidate_count
    dtype: float64
  - name: inference_calls_used
    dtype: float64
  - name: validation_calls_used
    dtype: float64
  - name: model_request_errors
    dtype: float64
  - name: elapsed_seconds
    dtype: float64
  - name: model_request_seconds
    dtype: float64
  - name: input_tokens
    dtype: float64
  - name: output_tokens
    dtype: float64
  - name: total_tokens
    dtype: float64
  - name: inference_cost_usd
    dtype: float64
  - name: compute_cost_usd
    dtype: float64
  - name: total_cost_usd
    dtype: float64
  - name: authoring_protocol
    dtype: string
  - name: max_output_tokens
    dtype: int64
  - name: output_limit
    dtype: string
  - name: reasoning_effort
    dtype: string
  - name: run_id
    dtype: string
  data_files:
  - split: nv25
    path:
    - viewer/nv25-gpt-6.1-sol-high-pi-none-turn64-20261020/scores.parquet
    - viewer/nv25-qwen38-27b-medium-pi-none-turn64-20261021/scores.parquet
    - viewer/nv25-qwen38-27b-xhigh-pi-none-turn64-20261021/scores.parquet
- config_name: by_protocol
  features:
  - name: reference_image
    dtype: image
  - name: candidate_image
    dtype: image
  - name: capture_view
    dtype: string
  - name: condition_id
    dtype: string
  - name: asset_id
    dtype: string
  - name: requested_model
    dtype: string
  - name: harness_id
    dtype: string
  - name: skill_set_id
    dtype: string
  - name: repetition
    dtype: int64
  - name: status
    dtype: string
  - name: formal_fidelity_percent
    dtype: float64
  - name: visual_fidelity_percent
    dtype: float64
  - name: kinematic_fidelity_percent
    dtype: float64
  - name: simready_reward
    dtype: float64
  - name: mandatory_passed
    dtype: float64
  - name: mandatory_total
    dtype: float64
  - name: mandatory_failed
    dtype: float64
  - name: mandatory_unknown
    dtype: float64
  - name: mandatory_lower_bound_percent
    dtype: float64
  - name: mandatory_upper_bound_percent
    dtype: float64
  - name: completion_bonus
    dtype: float64
  - name: optional_score_percent
    dtype: float64
  - name: optional_passed
    dtype: float64
  - name: optional_failed
    dtype: float64
  - name: optional_unknown
    dtype: float64
  - name: optional_total
    dtype: float64
  - name: optional_measured
    dtype: float64
  - name: optional_coverage
    dtype: float64
  - name: optional_lower_bound_percent
    dtype: float64
  - name: optional_upper_bound_percent
    dtype: float64
  - name: fidelity_shape_percent
    dtype: float64
  - name: fidelity_rigid_parts_percent
    dtype: float64
  - name: fidelity_joints_percent
    dtype: float64
  - name: fidelity_materials_percent
    dtype: float64
  - name: fidelity_shape_coverage
    dtype: float64
  - name: fidelity_rigid_parts_coverage
    dtype: float64
  - name: fidelity_joints_coverage
    dtype: float64
  - name: fidelity_materials_coverage
    dtype: float64
  - name: fidelity_reference_geometry_status
    dtype: string
  - name: formal_fidelity_status
    dtype: string
  - name: formal_fidelity_earned
    dtype: float64
  - name: formal_fidelity_eligible
    dtype: float64
  - name: formal_fidelity_measured
    dtype: float64
  - name: formal_fidelity_coverage
    dtype: float64
  - name: formal_fidelity_lower_bound_percent
    dtype: float64
  - name: formal_fidelity_upper_bound_percent
    dtype: float64
  - name: visual_fidelity_status
    dtype: string
  - name: visual_fidelity_earned
    dtype: float64
  - name: visual_fidelity_eligible
    dtype: float64
  - name: visual_fidelity_measured
    dtype: float64
  - name: visual_fidelity_coverage
    dtype: float64
  - name: visual_fidelity_lower_bound_percent
    dtype: float64
  - name: visual_fidelity_upper_bound_percent
    dtype: float64
  - name: kinematic_fidelity_status
    dtype: string
  - name: kinematic_fidelity_earned
    dtype: float64
  - name: kinematic_fidelity_eligible
    dtype: float64
  - name: kinematic_fidelity_measured
    dtype: float64
  - name: kinematic_fidelity_coverage
    dtype: float64
  - name: kinematic_fidelity_lower_bound_percent
    dtype: float64
  - name: kinematic_fidelity_upper_bound_percent
    dtype: float64
  - name: visual_appearance_percent
    dtype: float64
  - name: visual_appearance_status
    dtype: string
  - name: visual_appearance_earned
    dtype: float64
  - name: visual_appearance_eligible
    dtype: float64
  - name: visual_appearance_measured
    dtype: float64
  - name: visual_appearance_coverage
    dtype: float64
  - name: visual_appearance_lower_bound_percent
    dtype: float64
  - name: visual_appearance_upper_bound_percent
    dtype: float64
  - name: kinematic_reference_joint_count
    dtype: float64
  - name: kinematic_candidate_joint_count
    dtype: float64
  - name: kinematic_matched_joint_count
    dtype: float64
  - name: kinematic_missing_joint_count
    dtype: float64
  - name: kinematic_extra_joint_count
    dtype: float64
  - name: kinematic_moving_joints_percent
    dtype: float64
  - name: kinematic_moving_joints_status
    dtype: string
  - name: kinematic_moving_joints_earned
    dtype: float64
  - name: kinematic_moving_joints_eligible
    dtype: float64
  - name: kinematic_moving_joints_measured
    dtype: float64
  - name: kinematic_moving_joints_coverage
    dtype: float64
  - name: kinematic_moving_joints_lower_bound_percent
    dtype: float64
  - name: kinematic_moving_joints_upper_bound_percent
    dtype: float64
  - name: kinematic_fixed_constraints_percent
    dtype: float64
  - name: kinematic_fixed_constraints_status
    dtype: string
  - name: kinematic_fixed_constraints_earned
    dtype: float64
  - name: kinematic_fixed_constraints_eligible
    dtype: float64
  - name: kinematic_fixed_constraints_measured
    dtype: float64
  - name: kinematic_fixed_constraints_coverage
    dtype: float64
  - name: kinematic_fixed_constraints_lower_bound_percent
    dtype: float64
  - name: kinematic_fixed_constraints_upper_bound_percent
    dtype: float64
  - name: fidelity_shape_earned
    dtype: float64
  - name: fidelity_shape_eligible
    dtype: float64
  - name: fidelity_shape_measured
    dtype: float64
  - name: fidelity_shape_status
    dtype: string
  - name: fidelity_rigid_parts_earned
    dtype: float64
  - name: fidelity_rigid_parts_eligible
    dtype: float64
  - name: fidelity_rigid_parts_measured
    dtype: float64
  - name: fidelity_rigid_parts_status
    dtype: string
  - name: fidelity_joints_earned
    dtype: float64
  - name: fidelity_joints_eligible
    dtype: float64
  - name: fidelity_joints_measured
    dtype: float64
  - name: fidelity_joints_status
    dtype: string
  - name: fidelity_materials_earned
    dtype: float64
  - name: fidelity_materials_eligible
    dtype: float64
  - name: fidelity_materials_measured
    dtype: float64
  - name: fidelity_materials_status
    dtype: string
  - name: fidelity_shape_precision
    dtype: float64
  - name: fidelity_shape_recall
    dtype: float64
  - name: fidelity_shape_samples_each_direction
    dtype: float64
  - name: fidelity_rigid_parts_geometry_percent
    dtype: float64
  - name: fidelity_rigid_parts_structure_percent
    dtype: float64
  - name: fidelity_rigid_parts_structure_earned
    dtype: float64
  - name: fidelity_rigid_parts_structure_eligible
    dtype: float64
  - name: fidelity_joints_reference_count
    dtype: float64
  - name: fidelity_joints_candidate_count
    dtype: float64
  - name: inference_calls_used
    dtype: float64
  - name: validation_calls_used
    dtype: float64
  - name: model_request_errors
    dtype: float64
  - name: elapsed_seconds
    dtype: float64
  - name: model_request_seconds
    dtype: float64
  - name: input_tokens
    dtype: float64
  - name: output_tokens
    dtype: float64
  - name: total_tokens
    dtype: float64
  - name: inference_cost_usd
    dtype: float64
  - name: compute_cost_usd
    dtype: float64
  - name: total_cost_usd
    dtype: float64
  - name: authoring_protocol
    dtype: string
  - name: max_output_tokens
    dtype: int64
  - name: output_limit
    dtype: string
  - name: reasoning_effort
    dtype: string
  - name: run_id
    dtype: string
  data_files:
  - split: authoring_64_bounded_output_8192_v1
    path:
    - viewer/nv25-qwen38-27b-medium-pi-none-turn64-20261021/scores.parquet
  - split: authoring_64_native_output_v1
    path:
    - viewer/nv25-gpt-6.1-sol-high-pi-none-turn64-20261020/scores.parquet
    - viewer/nv25-qwen38-27b-xhigh-pi-none-turn64-20261021/scores.parquet
- config_name: by_reasoning
  features:
  - name: reference_image
    dtype: image
  - name: candidate_image
    dtype: image
  - name: capture_view
    dtype: string
  - name: condition_id
    dtype: string
  - name: asset_id
    dtype: string
  - name: requested_model
    dtype: string
  - name: harness_id
    dtype: string
  - name: skill_set_id
    dtype: string
  - name: repetition
    dtype: int64
  - name: status
    dtype: string
  - name: formal_fidelity_percent
    dtype: float64
  - name: visual_fidelity_percent
    dtype: float64
  - name: kinematic_fidelity_percent
    dtype: float64
  - name: simready_reward
    dtype: float64
  - name: mandatory_passed
    dtype: float64
  - name: mandatory_total
    dtype: float64
  - name: mandatory_failed
    dtype: float64
  - name: mandatory_unknown
    dtype: float64
  - name: mandatory_lower_bound_percent
    dtype: float64
  - name: mandatory_upper_bound_percent
    dtype: float64
  - name: completion_bonus
    dtype: float64
  - name: optional_score_percent
    dtype: float64
  - name: optional_passed
    dtype: float64
  - name: optional_failed
    dtype: float64
  - name: optional_unknown
    dtype: float64
  - name: optional_total
    dtype: float64
  - name: optional_measured
    dtype: float64
  - name: optional_coverage
    dtype: float64
  - name: optional_lower_bound_percent
    dtype: float64
  - name: optional_upper_bound_percent
    dtype: float64
  - name: fidelity_shape_percent
    dtype: float64
  - name: fidelity_rigid_parts_percent
    dtype: float64
  - name: fidelity_joints_percent
    dtype: float64
  - name: fidelity_materials_percent
    dtype: float64
  - name: fidelity_shape_coverage
    dtype: float64
  - name: fidelity_rigid_parts_coverage
    dtype: float64
  - name: fidelity_joints_coverage
    dtype: float64
  - name: fidelity_materials_coverage
    dtype: float64
  - name: fidelity_reference_geometry_status
    dtype: string
  - name: formal_fidelity_status
    dtype: string
  - name: formal_fidelity_earned
    dtype: float64
  - name: formal_fidelity_eligible
    dtype: float64
  - name: formal_fidelity_measured
    dtype: float64
  - name: formal_fidelity_coverage
    dtype: float64
  - name: formal_fidelity_lower_bound_percent
    dtype: float64
  - name: formal_fidelity_upper_bound_percent
    dtype: float64
  - name: visual_fidelity_status
    dtype: string
  - name: visual_fidelity_earned
    dtype: float64
  - name: visual_fidelity_eligible
    dtype: float64
  - name: visual_fidelity_measured
    dtype: float64
  - name: visual_fidelity_coverage
    dtype: float64
  - name: visual_fidelity_lower_bound_percent
    dtype: float64
  - name: visual_fidelity_upper_bound_percent
    dtype: float64
  - name: kinematic_fidelity_status
    dtype: string
  - name: kinematic_fidelity_earned
    dtype: float64
  - name: kinematic_fidelity_eligible
    dtype: float64
  - name: kinematic_fidelity_measured
    dtype: float64
  - name: kinematic_fidelity_coverage
    dtype: float64
  - name: kinematic_fidelity_lower_bound_percent
    dtype: float64
  - name: kinematic_fidelity_upper_bound_percent
    dtype: float64
  - name: visual_appearance_percent
    dtype: float64
  - name: visual_appearance_status
    dtype: string
  - name: visual_appearance_earned
    dtype: float64
  - name: visual_appearance_eligible
    dtype: float64
  - name: visual_appearance_measured
    dtype: float64
  - name: visual_appearance_coverage
    dtype: float64
  - name: visual_appearance_lower_bound_percent
    dtype: float64
  - name: visual_appearance_upper_bound_percent
    dtype: float64
  - name: kinematic_reference_joint_count
    dtype: float64
  - name: kinematic_candidate_joint_count
    dtype: float64
  - name: kinematic_matched_joint_count
    dtype: float64
  - name: kinematic_missing_joint_count
    dtype: float64
  - name: kinematic_extra_joint_count
    dtype: float64
  - name: kinematic_moving_joints_percent
    dtype: float64
  - name: kinematic_moving_joints_status
    dtype: string
  - name: kinematic_moving_joints_earned
    dtype: float64
  - name: kinematic_moving_joints_eligible
    dtype: float64
  - name: kinematic_moving_joints_measured
    dtype: float64
  - name: kinematic_moving_joints_coverage
    dtype: float64
  - name: kinematic_moving_joints_lower_bound_percent
    dtype: float64
  - name: kinematic_moving_joints_upper_bound_percent
    dtype: float64
  - name: kinematic_fixed_constraints_percent
    dtype: float64
  - name: kinematic_fixed_constraints_status
    dtype: string
  - name: kinematic_fixed_constraints_earned
    dtype: float64
  - name: kinematic_fixed_constraints_eligible
    dtype: float64
  - name: kinematic_fixed_constraints_measured
    dtype: float64
  - name: kinematic_fixed_constraints_coverage
    dtype: float64
  - name: kinematic_fixed_constraints_lower_bound_percent
    dtype: float64
  - name: kinematic_fixed_constraints_upper_bound_percent
    dtype: float64
  - name: fidelity_shape_earned
    dtype: float64
  - name: fidelity_shape_eligible
    dtype: float64
  - name: fidelity_shape_measured
    dtype: float64
  - name: fidelity_shape_status
    dtype: string
  - name: fidelity_rigid_parts_earned
    dtype: float64
  - name: fidelity_rigid_parts_eligible
    dtype: float64
  - name: fidelity_rigid_parts_measured
    dtype: float64
  - name: fidelity_rigid_parts_status
    dtype: string
  - name: fidelity_joints_earned
    dtype: float64
  - name: fidelity_joints_eligible
    dtype: float64
  - name: fidelity_joints_measured
    dtype: float64
  - name: fidelity_joints_status
    dtype: string
  - name: fidelity_materials_earned
    dtype: float64
  - name: fidelity_materials_eligible
    dtype: float64
  - name: fidelity_materials_measured
    dtype: float64
  - name: fidelity_materials_status
    dtype: string
  - name: fidelity_shape_precision
    dtype: float64
  - name: fidelity_shape_recall
    dtype: float64
  - name: fidelity_shape_samples_each_direction
    dtype: float64
  - name: fidelity_rigid_parts_geometry_percent
    dtype: float64
  - name: fidelity_rigid_parts_structure_percent
    dtype: float64
  - name: fidelity_rigid_parts_structure_earned
    dtype: float64
  - name: fidelity_rigid_parts_structure_eligible
    dtype: float64
  - name: fidelity_joints_reference_count
    dtype: float64
  - name: fidelity_joints_candidate_count
    dtype: float64
  - name: inference_calls_used
    dtype: float64
  - name: validation_calls_used
    dtype: float64
  - name: model_request_errors
    dtype: float64
  - name: elapsed_seconds
    dtype: float64
  - name: model_request_seconds
    dtype: float64
  - name: input_tokens
    dtype: float64
  - name: output_tokens
    dtype: float64
  - name: total_tokens
    dtype: float64
  - name: inference_cost_usd
    dtype: float64
  - name: compute_cost_usd
    dtype: float64
  - name: total_cost_usd
    dtype: float64
  - name: authoring_protocol
    dtype: string
  - name: max_output_tokens
    dtype: int64
  - name: output_limit
    dtype: string
  - name: reasoning_effort
    dtype: string
  - name: run_id
    dtype: string
  data_files:
  - split: high
    path:
    - viewer/nv25-gpt-6.1-sol-high-pi-none-turn64-20261020/scores.parquet
  - split: medium
    path:
    - viewer/nv25-qwen38-27b-medium-pi-none-turn64-20261021/scores.parquet
  - split: xhigh
    path:
    - viewer/nv25-qwen38-27b-xhigh-pi-none-turn64-20261021/scores.parquet
- config_name: by_asset
  features:
  - name: reference_image
    dtype: image
  - name: candidate_image
    dtype: image
  - name: capture_view
    dtype: string
  - name: condition_id
    dtype: string
  - name: asset_id
    dtype: string
  - name: requested_model
    dtype: string
  - name: harness_id
    dtype: string
  - name: skill_set_id
    dtype: string
  - name: repetition
    dtype: int64
  - name: status
    dtype: string
  - name: formal_fidelity_percent
    dtype: float64
  - name: visual_fidelity_percent
    dtype: float64
  - name: kinematic_fidelity_percent
    dtype: float64
  - name: simready_reward
    dtype: float64
  - name: mandatory_passed
    dtype: float64
  - name: mandatory_total
    dtype: float64
  - name: mandatory_failed
    dtype: float64
  - name: mandatory_unknown
    dtype: float64
  - name: mandatory_lower_bound_percent
    dtype: float64
  - name: mandatory_upper_bound_percent
    dtype: float64
  - name: completion_bonus
    dtype: float64
  - name: optional_score_percent
    dtype: float64
  - name: optional_passed
    dtype: float64
  - name: optional_failed
    dtype: float64
  - name: optional_unknown
    dtype: float64
  - name: optional_total
    dtype: float64
  - name: optional_measured
    dtype: float64
  - name: optional_coverage
    dtype: float64
  - name: optional_lower_bound_percent
    dtype: float64
  - name: optional_upper_bound_percent
    dtype: float64
  - name: fidelity_shape_percent
    dtype: float64
  - name: fidelity_rigid_parts_percent
    dtype: float64
  - name: fidelity_joints_percent
    dtype: float64
  - name: fidelity_materials_percent
    dtype: float64
  - name: fidelity_shape_coverage
    dtype: float64
  - name: fidelity_rigid_parts_coverage
    dtype: float64
  - name: fidelity_joints_coverage
    dtype: float64
  - name: fidelity_materials_coverage
    dtype: float64
  - name: fidelity_reference_geometry_status
    dtype: string
  - name: formal_fidelity_status
    dtype: string
  - name: formal_fidelity_earned
    dtype: float64
  - name: formal_fidelity_eligible
    dtype: float64
  - name: formal_fidelity_measured
    dtype: float64
  - name: formal_fidelity_coverage
    dtype: float64
  - name: formal_fidelity_lower_bound_percent
    dtype: float64
  - name: formal_fidelity_upper_bound_percent
    dtype: float64
  - name: visual_fidelity_status
    dtype: string
  - name: visual_fidelity_earned
    dtype: float64
  - name: visual_fidelity_eligible
    dtype: float64
  - name: visual_fidelity_measured
    dtype: float64
  - name: visual_fidelity_coverage
    dtype: float64
  - name: visual_fidelity_lower_bound_percent
    dtype: float64
  - name: visual_fidelity_upper_bound_percent
    dtype: float64
  - name: kinematic_fidelity_status
    dtype: string
  - name: kinematic_fidelity_earned
    dtype: float64
  - name: kinematic_fidelity_eligible
    dtype: float64
  - name: kinematic_fidelity_measured
    dtype: float64
  - name: kinematic_fidelity_coverage
    dtype: float64
  - name: kinematic_fidelity_lower_bound_percent
    dtype: float64
  - name: kinematic_fidelity_upper_bound_percent
    dtype: float64
  - name: visual_appearance_percent
    dtype: float64
  - name: visual_appearance_status
    dtype: string
  - name: visual_appearance_earned
    dtype: float64
  - name: visual_appearance_eligible
    dtype: float64
  - name: visual_appearance_measured
    dtype: float64
  - name: visual_appearance_coverage
    dtype: float64
  - name: visual_appearance_lower_bound_percent
    dtype: float64
  - name: visual_appearance_upper_bound_percent
    dtype: float64
  - name: kinematic_reference_joint_count
    dtype: float64
  - name: kinematic_candidate_joint_count
    dtype: float64
  - name: kinematic_matched_joint_count
    dtype: float64
  - name: kinematic_missing_joint_count
    dtype: float64
  - name: kinematic_extra_joint_count
    dtype: float64
  - name: kinematic_moving_joints_percent
    dtype: float64
  - name: kinematic_moving_joints_status
    dtype: string
  - name: kinematic_moving_joints_earned
    dtype: float64
  - name: kinematic_moving_joints_eligible
    dtype: float64
  - name: kinematic_moving_joints_measured
    dtype: float64
  - name: kinematic_moving_joints_coverage
    dtype: float64
  - name: kinematic_moving_joints_lower_bound_percent
    dtype: float64
  - name: kinematic_moving_joints_upper_bound_percent
    dtype: float64
  - name: kinematic_fixed_constraints_percent
    dtype: float64
  - name: kinematic_fixed_constraints_status
    dtype: string
  - name: kinematic_fixed_constraints_earned
    dtype: float64
  - name: kinematic_fixed_constraints_eligible
    dtype: float64
  - name: kinematic_fixed_constraints_measured
    dtype: float64
  - name: kinematic_fixed_constraints_coverage
    dtype: float64
  - name: kinematic_fixed_constraints_lower_bound_percent
    dtype: float64
  - name: kinematic_fixed_constraints_upper_bound_percent
    dtype: float64
  - name: fidelity_shape_earned
    dtype: float64
  - name: fidelity_shape_eligible
    dtype: float64
  - name: fidelity_shape_measured
    dtype: float64
  - name: fidelity_shape_status
    dtype: string
  - name: fidelity_rigid_parts_earned
    dtype: float64
  - name: fidelity_rigid_parts_eligible
    dtype: float64
  - name: fidelity_rigid_parts_measured
    dtype: float64
  - name: fidelity_rigid_parts_status
    dtype: string
  - name: fidelity_joints_earned
    dtype: float64
  - name: fidelity_joints_eligible
    dtype: float64
  - name: fidelity_joints_measured
    dtype: float64
  - name: fidelity_joints_status
    dtype: string
  - name: fidelity_materials_earned
    dtype: float64
  - name: fidelity_materials_eligible
    dtype: float64
  - name: fidelity_materials_measured
    dtype: float64
  - name: fidelity_materials_status
    dtype: string
  - name: fidelity_shape_precision
    dtype: float64
  - name: fidelity_shape_recall
    dtype: float64
  - name: fidelity_shape_samples_each_direction
    dtype: float64
  - name: fidelity_rigid_parts_geometry_percent
    dtype: float64
  - name: fidelity_rigid_parts_structure_percent
    dtype: float64
  - name: fidelity_rigid_parts_structure_earned
    dtype: float64
  - name: fidelity_rigid_parts_structure_eligible
    dtype: float64
  - name: fidelity_joints_reference_count
    dtype: float64
  - name: fidelity_joints_candidate_count
    dtype: float64
  - name: inference_calls_used
    dtype: float64
  - name: validation_calls_used
    dtype: float64
  - name: model_request_errors
    dtype: float64
  - name: elapsed_seconds
    dtype: float64
  - name: model_request_seconds
    dtype: float64
  - name: input_tokens
    dtype: float64
  - name: output_tokens
    dtype: float64
  - name: total_tokens
    dtype: float64
  - name: inference_cost_usd
    dtype: float64
  - name: compute_cost_usd
    dtype: float64
  - name: total_cost_usd
    dtype: float64
  - name: authoring_protocol
    dtype: string
  - name: max_output_tokens
    dtype: int64
  - name: output_limit
    dtype: string
  - name: reasoning_effort
    dtype: string
  - name: run_id
    dtype: string
  data_files:
  - split: accessory_eyeglasses_a01
    path:
    - viewer/nv25-gpt-6.1-sol-high-pi-none-turn64-20261020/by_asset/accessory_eyeglasses_a01.parquet
    - viewer/nv25-qwen38-27b-medium-pi-none-turn64-20261021/by_asset/accessory_eyeglasses_a01.parquet
    - viewer/nv25-qwen38-27b-xhigh-pi-none-turn64-20261021/by_asset/accessory_eyeglasses_a01.parquet
  - split: box_cardboard_open_a01
    path:
    - viewer/nv25-gpt-6.1-sol-high-pi-none-turn64-20261020/by_asset/box_cardboard_open_a01.parquet
    - viewer/nv25-qwen38-27b-medium-pi-none-turn64-20261021/by_asset/box_cardboard_open_a01.parquet
  - split: bucket_utility_red_a01
    path:
    - viewer/nv25-gpt-6.1-sol-high-pi-none-turn64-20261020/by_asset/bucket_utility_red_a01.parquet
  - split: chair_office_amf_black_a01
    path:
    - viewer/nv25-gpt-6.1-sol-high-pi-none-turn64-20261020/by_asset/chair_office_amf_black_a01.parquet
  - split: cleaning_cloth_a01
    path:
    - viewer/nv25-gpt-6.1-sol-high-pi-none-turn64-20261020/by_asset/cleaning_cloth_a01.parquet
    - viewer/nv25-qwen38-27b-medium-pi-none-turn64-20261021/by_asset/cleaning_cloth_a01.parquet
  - split: container_bottle_glass_a01
    path:
    - viewer/nv25-gpt-6.1-sol-high-pi-none-turn64-20261020/by_asset/container_bottle_glass_a01.parquet
    - viewer/nv25-qwen38-27b-medium-pi-none-turn64-20261021/by_asset/container_bottle_glass_a01.parquet
  - split: enginecomponent_camshaft_a01
    path:
    - viewer/nv25-gpt-6.1-sol-high-pi-none-turn64-20261020/by_asset/enginecomponent_camshaft_a01.parquet
    - viewer/nv25-qwen38-27b-medium-pi-none-turn64-20261021/by_asset/enginecomponent_camshaft_a01.parquet
  - split: food_beverage_coffeecup_a01
    path:
    - viewer/nv25-gpt-6.1-sol-high-pi-none-turn64-20261020/by_asset/food_beverage_coffeecup_a01.parquet
  - split: food_prepared_croissant_a01
    path:
    - viewer/nv25-gpt-6.1-sol-high-pi-none-turn64-20261020/by_asset/food_prepared_croissant_a01.parquet
    - viewer/nv25-qwen38-27b-medium-pi-none-turn64-20261021/by_asset/food_prepared_croissant_a01.parquet
  - split: food_produce_banana_a01
    path:
    - viewer/nv25-gpt-6.1-sol-high-pi-none-turn64-20261020/by_asset/food_produce_banana_a01.parquet
    - viewer/nv25-qwen38-27b-medium-pi-none-turn64-20261021/by_asset/food_produce_banana_a01.parquet
  - split: food_produce_broccoli_b01
    path:
    - viewer/nv25-gpt-6.1-sol-high-pi-none-turn64-20261020/by_asset/food_produce_broccoli_b01.parquet
    - viewer/nv25-qwen38-27b-medium-pi-none-turn64-20261021/by_asset/food_produce_broccoli_b01.parquet
    - viewer/nv25-qwen38-27b-xhigh-pi-none-turn64-20261021/by_asset/food_produce_broccoli_b01.parquet
  - split: handtool_hammer_crosspeen_a01
    path:
    - viewer/nv25-gpt-6.1-sol-high-pi-none-turn64-20261020/by_asset/handtool_hammer_crosspeen_a01.parquet
  - split: handtool_wrench_adjustable_a01
    path:
    - viewer/nv25-gpt-6.1-sol-high-pi-none-turn64-20261020/by_asset/handtool_wrench_adjustable_a01.parquet
  - split: handtruck_convertible_steel_red_a01
    path:
    - viewer/nv25-gpt-6.1-sol-high-pi-none-turn64-20261020/by_asset/handtruck_convertible_steel_red_a01.parquet
  - split: kitchenware_kettle_a01
    path:
    - viewer/nv25-gpt-6.1-sol-high-pi-none-turn64-20261020/by_asset/kitchenware_kettle_a01.parquet
  - split: kitchenware_pan_j01
    path:
    - viewer/nv25-gpt-6.1-sol-high-pi-none-turn64-20261020/by_asset/kitchenware_pan_j01.parquet
    - viewer/nv25-qwen38-27b-medium-pi-none-turn64-20261021/by_asset/kitchenware_pan_j01.parquet
  - split: pallet_wood_recycled_a05
    path:
    - viewer/nv25-gpt-6.1-sol-high-pi-none-turn64-20261020/by_asset/pallet_wood_recycled_a05.parquet
    - viewer/nv25-qwen38-27b-medium-pi-none-turn64-20261021/by_asset/pallet_wood_recycled_a05.parquet
  - split: powertool_drill_b01
    path:
    - viewer/nv25-gpt-6.1-sol-high-pi-none-turn64-20261020/by_asset/powertool_drill_b01.parquet
  - split: ppe_helmet_hardhat_a01
    path:
    - viewer/nv25-gpt-6.1-sol-high-pi-none-turn64-20261020/by_asset/ppe_helmet_hardhat_a01.parquet
    - viewer/nv25-qwen38-27b-medium-pi-none-turn64-20261021/by_asset/ppe_helmet_hardhat_a01.parquet
  - split: servicetool_funnel_flexiblespout_a01
    path:
    - viewer/nv25-gpt-6.1-sol-high-pi-none-turn64-20261020/by_asset/servicetool_funnel_flexiblespout_a01.parquet
  - split: sign_wetfloor_a01
    path:
    - viewer/nv25-gpt-6.1-sol-high-pi-none-turn64-20261020/by_asset/sign_wetfloor_a01.parquet
  - split: suspension_coilover_rear_b01
    path:
    - viewer/nv25-gpt-6.1-sol-high-pi-none-turn64-20261020/by_asset/suspension_coilover_rear_b01.parquet
    - viewer/nv25-qwen38-27b-medium-pi-none-turn64-20261021/by_asset/suspension_coilover_rear_b01.parquet
  - split: tools_fireextinguisher_co2_a01
    path:
    - viewer/nv25-gpt-6.1-sol-high-pi-none-turn64-20261020/by_asset/tools_fireextinguisher_co2_a01.parquet
  - split: utensils_scissors_a01
    path:
    - viewer/nv25-gpt-6.1-sol-high-pi-none-turn64-20261020/by_asset/utensils_scissors_a01.parquet
    - viewer/nv25-qwen38-27b-medium-pi-none-turn64-20261021/by_asset/utensils_scissors_a01.parquet
  - split: wheelassembly_wheel_rim_a01
    path:
    - viewer/nv25-gpt-6.1-sol-high-pi-none-turn64-20261020/by_asset/wheelassembly_wheel_rim_a01.parquet
    - viewer/nv25-qwen38-27b-medium-pi-none-turn64-20261021/by_asset/wheelassembly_wheel_rim_a01.parquet
version: 1.0.0
simreadybench_viewer: 3
tags:
- benchmark
- evaluation
- results
- 3d
- robotics
- simready
task_categories:
- image-to-3d
- text-to-3d
---
<div align="center">
  <h1>SimReadyBench</h1>
  <p><strong>A benchmark for simulation-ready asset reconstruction</strong></p>
  <p>
    <a href="https://chrisvoncsefalvay.com">Chris von Csefalvay</a>,
    Pranav Doma, Achyuthan Unni Krishnan and
    <a href="https://github.com/tfoldi">Tamas Foldi</a>
  </p>
  <p><a href="https://www.hcltech.com/">HCLTech</a></p>
  <p>
    <a href="https://hcltech-robotics.github.io/simreadybench/"><img src="https://img.shields.io/badge/Project-page-006B73?style=flat-square" alt="Project page"></a>
    <a href="https://hcltech-robotics.github.io/simreadybench/#paper"><img src="https://img.shields.io/badge/Paper-coming_soon-8A6D3B?style=flat-square" alt="Paper coming soon"></a>
    <a href="https://github.com/hcltech-robotics/simreadybench"><img src="https://img.shields.io/badge/Code-GitHub-24292F?style=flat-square" alt="Code on GitHub"></a>
    <a href="https://huggingface.co/datasets/hcltech-robotics/simreadybench"><img src="https://img.shields.io/badge/Dataset-results-567A9A?style=flat-square" alt="Dataset and results on Hugging Face"></a>
    <a href="https://github.com/hcltech-robotics/simreadybench/blob/main/docs/evaluation.md"><img src="https://img.shields.io/badge/Scoring-guide-5B6573?style=flat-square" alt="Scoring guide"></a>
  </p>
  <p>
    <a href="https://huggingface.co/datasets/hcltech-robotics/simreadybench/blob/main/CODE-LICENSE.txt"><img src="https://img.shields.io/badge/Code-MIT-5B6573?style=flat-square" alt="Code: MIT licence"></a>
    <a href="https://huggingface.co/datasets/hcltech-robotics/simreadybench/blob/main/NVIDIA-ASSET-LICENSE.txt"><img src="https://img.shields.io/badge/Source_assets-CC_BY_4.0-5B6573?style=flat-square" alt="Source assets: CC BY 4.0 licence"></a>
  </p>
</div>

SimReadyBench measures how well models reconstruct simulation-ready USD assets from reference images, an authoring brief and nominal dimensions. Models, agent harnesses and authoring skills are independent experimental factors. Evaluation reports three independent composite scores: formal fidelity, visual fidelity and kinematic fidelity.

<!-- simreadybench-leaderboard:start -->
## NV25 leaderboard

Each entry contains 100 reconstructions: four attempts for each of the 25 NV25 assets. Formal, visual and kinematic fidelity are independent scores on a 0-100 scale, with no combined ranking. Optional results score the 51 optional SimReady requirements separately from the 46 mandatory requirements in formal fidelity. Values are means [95% confidence interval].

**NV25, Robotics-Prop 4.0.0**. Protocol: `authoring-64-native-output-v1`. Reference revision: `f4860aa0c75607adb50a78bf96e77510bffc040f`.

| Model | Reasoning | Harness | Skills | Formal (%) | Optional (%) | Visual (%) | Kinematic (%) | Results |
| --- | --- | --- | --- | ---: | ---: | ---: | ---: | --- |
| gpt-6.1-sol | high | pi | none | 100.00 [100.00, 100.00] | 73.78 [72.96, 74.57] | 44.87 [37.01, 52.51] | 7.04 [0.24, 19.84] | <a href="https://huggingface.co/datasets/hcltech-robotics/simreadybench/blob/main/runs/nv25-gpt-6.1-sol-high-pi-none-turn64-20261020/report.html">Readout</a> |

Confidence intervals use 10,000 asset-cluster bootstrap samples, retaining the four reconstructions of each sampled asset together. N/A excludes references without eligible criteria. Unknown measurements remain unknown. [Scores, coverage and source revisions](leaderboard.json).

![Formal, visual and kinematic fidelity means with 95% confidence intervals.](media/leaderboard-scores.png)

### Efficiency

Efficiency is the mean fidelity divided by mean resource use across eligible reconstructions with both measurements. Token efficiency uses total input and output tokens, while time efficiency uses elapsed reconstruction time, including model requests, tools and evaluation. Cost efficiency uses pricing-derived inference cost in USD, with compute cost reported separately. Each ratio uses paired score and resource measurements, with joint asset-cluster bootstrap intervals.

| Configuration | Formal points/1M tokens | Visual points/1M tokens | Kinematic points/1M tokens |
| --- | ---: | ---: | ---: |
| gpt-6.1-sol<br>high reasoning, pi, skills: none<br>authoring-64-native-output-v1 | 102.90 [95.98, 110.80]* | 46.34 [37.17, 55.88]* | 6.83 [0.25, 19.09]* |

![Formal, visual and kinematic fidelity points per million tokens with 95% confidence intervals.](media/leaderboard-token-efficiency.png)

| Configuration | Formal points/minute | Visual points/minute | Kinematic points/minute |
| --- | ---: | ---: | ---: |
| gpt-6.1-sol<br>high reasoning, pi, skills: none<br>authoring-64-native-output-v1 | 9.70 [9.13, 10.30] | 4.35 [3.48, 5.25] | 0.65 [0.02, 1.86] |

![Formal, visual and kinematic fidelity points per elapsed minute with 95% confidence intervals.](media/leaderboard-time-efficiency.png)

| Configuration | Formal points/USD inference | Visual points/USD inference | Kinematic points/USD inference |
| --- | ---: | ---: | ---: |
| gpt-6.1-sol<br>high reasoning, pi, skills: none<br>authoring-64-native-output-v1 | 219.61 [204.88, 234.78]* | 98.90 [79.01, 118.91]* | 14.48 [0.51, 41.55]* |

![Formal, visual and kinematic fidelity points per USD of pricing-derived inference cost with 95% confidence intervals.](media/leaderboard-inference-cost-efficiency.png)

### Resource use

Means per reconstruction with measured usage. Full precision, totals and measurement coverage are retained in [leaderboard.json](leaderboard.json).

| Configuration | Total tokens | Elapsed minutes | Model requests | Pricing-derived inference cost (USD) | Compute cost (USD) | Total cost (USD) |
| --- | ---: | ---: | ---: | ---: | ---: | ---: |
| gpt-6.1-sol<br>high reasoning, pi, skills: none<br>authoring-64-native-output-v1 | 971,775.96* | 10.31 | 36.02 | 0.46* | unknown | unknown |

\* Incomplete usage coverage.

### Strict conformance

A strict pass satisfies all 46 mandatory SimReady requirements. The four-attempt rate is the fraction of assets with at least one strict pass among their four reconstructions.

| Configuration | Strict pass per reconstruction | At least one strict pass in four |
| --- | ---: | ---: |
| gpt-6.1-sol<br>high reasoning, pi, skills: none<br>authoring-64-native-output-v1 | 100.00% (100/100) | 100.00% (25/25) |

<!-- simreadybench-leaderboard:end -->

## Reference sets

NV25 and NV100 are distinct public reference sets, separate from the withheld internal benchmark reference set. The packaged [`nvidia-25` manifest](https://github.com/hcltech-robotics/simreadybench/blob/main/src/simready_bench/resources/benchmarks/nvidia-25.json) defines NV25, which contains 25 NVIDIA SimReady assets.

## Evaluation protocol

![Evaluation protocol from a reference USD through controlled observations and reconstruction to profile validation, fidelity measurement and matched captures.](media/evaluation-protocol.png)

The NV25 protocol uses four independent reconstructions per asset, giving 100 attempts for each model, harness and skill configuration. Reference observations and eligible measurements are fixed before authoring. Each attempt records the model settings and reconstruction identity.

| Experimental factor | Definition |
| --- | --- |
| Public reference set | NV25, with the same 25 assets and their declared dependencies for every configuration |
| Model | The requested model with recorded generation and reasoning settings |
| Harness | The selected agent harness, such as Pi or OpenCode, used to author and submit the USD |
| Skills | The supplied authoring skill set, including an explicit empty condition |
| Repetitions | Four distinct reconstruction identities for each reference |

During a reconstruction, the agent receives reference views, dimensions and an authoring brief. It can inspect the asset observations and request validator feedback as it authors a self-contained USD package. Final evaluation uses the submitted package, with shared orientation-aligned cameras for reference and reconstruction captures.

## Evaluation factors

![An eyeglasses reference split between wireframe and material appearance, with callouts for shape, rigid parts, joints and materials.](media/evaluation-factors.png)

The three composites describe different properties of the reconstruction and are reported independently on a 0-100 scale:

| Composite | Factors | Report field |
| --- | --- | --- |
| Formal fidelity | Mandatory SimReady Robotics-Prop requirements and strict profile pass | `formal_fidelity_percent` |
| Visual fidelity | Reference-scale shape and appearance material correspondence | `visual_fidelity_percent` |
| Kinematic fidelity | Joint recovery, connectivity and correspondence of allowed relative poses | `kinematic_fidelity_percent` |

### Formal fidelity

<!-- score-diagram:formal:start -->

![Formal fidelity combines passed mandatory requirement credit with the all-pass bonus.](media/formal-fidelity.png)

[SVG](media/formal-fidelity.svg) | [Editable Mermaid source](media/formal-fidelity.mmd)

<!-- score-diagram:formal:end -->

Profile validation uses [NVIDIA SimReady-Foundation Robotics-Prop 4.0.0](https://github.com/NVIDIA/simready-foundation/blob/main/nv_core/sr_specs/docs/guides/profiles_validation_workflow.md). The 46 mandatory requirements cover the structural, physical and semantic properties of a simulation-ready prop. Each mandatory requirement contributes equally to 95 points, with five additional points when all 46 pass:

```text
Formal fidelity = 95 x passed mandatory requirements / 46
                + 5 when all 46 mandatory requirements pass
```

Harbor receives this value as `simready_reward`. Strict acceptance requires all 46 mandatory requirements to pass. Multiple findings against one requirement count as one failed requirement. Missing or invalid candidate output scores zero against the same 46 requirements, while an incomplete evaluator run has a null score with credit bounds.

The profile's 51 optional requirements receive a separate score:

```text
Optional score = 100 x passed optional requirements / 51
```

Both scores retain their passed, failed and unknown requirement counts. Partially measured profiles include coverage and lower and upper score bounds. Optional requirements do not change formal fidelity or its completion bonus.

### Visual fidelity

<!-- score-diagram:visual:start -->

![Visual fidelity is the geometric mean of shape and supported material appearance.](media/visual-fidelity.png)


<!-- score-diagram:visual:end -->

Visual fidelity is the geometric mean of shape and appearance correspondence:

```text
Visual fidelity = sqrt(shape_percent x appearance_percent)
```

Shape is the bidirectional surface F-score at the reference scale. Appearance compares supported colour, roughness, metallic response, opacity and texture values at corresponding reference surface samples. Each sample receives the mean of its eligible property matches, then the equal-area samples are averaged. The material-assignment flag, friction, restitution and density do not contribute to appearance credit. Assigned reference regions without measurable appearance inputs retain unknown credit, measurement coverage and lower and upper bounds. A factor with no eligible reference criteria is not applicable and is excluded from the mean.

### Kinematic fidelity

<!-- score-diagram:kinematic:start -->

![Kinematic fidelity combines per-joint agreement through moving-joint and fixed-constraint recovery.](media/kinematic-fidelity.png)

<!-- score-diagram:kinematic:end -->

The evaluator matches joints one to one through corresponding bodies and compares the relative poses permitted by their constraints. For fixed, revolute and prismatic joints, seventeen deterministic nonuniform states sample each motion probe. Both constraints act on the same reference-body landmarks, so reconstructed surface detail does not count twice as kinematic error.

```text
band = 0.001 m + 0.01 x reference bounding-box diagonal
motion credit = 1 / (1 + (motion_rms_m / band)^2)
joint quality = motion credit x range credit x bound-pattern credit
group score = 100 x 2 x sum(matched joint quality)
                  / (reference joint count + candidate joint count)
```

Finite ranges receive continuous credit for their unwrapped travel span, using `1 / (1 + (span_error / span_band)^2)` with a span band of five degrees or 5 mm plus 5% of the reference span. Bound-pattern credit is `1 / (1 + differing_endpoints)`, where an endpoint differs when one interval is finite and the other is infinite. Fixed constraints have unit range and bound-pattern credit.

Known unbounded revolute joints use a canonical probe spanning one full revolution and prismatic joints use one reference diagonal. A half-bounded probe is anchored at its finite endpoint, while a fully unbounded probe uses a canonical coordinate origin. Authored bounds remain unchanged in the measurement record. Missing, blocked or contradictory bound observations remain unknown.

Type, connectivity and the relevant enabled/kinematic states must agree for motion credit. Missing joints receive no credit and additional joints increase the denominator. Fixed attachments and moving joints form separate groups. Their geometric mean gives the kinematic composite when both groups apply, so fixed attachments cannot compensate for missing movable joints. Equivalent axis signs, reversed limits, exchanged joint endpoints and pivots shifted along the same hinge line retain the same score. Unsupported reference constraints and ambiguous body correspondence remain unknown. A reference with no joints is not applicable when the candidate also has none, while additional candidate joints score zero.

### Detailed correspondence measurements

Units and up-axis are normalised to metres/Z-up, then one whole-object translation and rotation about the vertical axis align the candidate with the reference. The same transform is applied to geometry, bodies, joint frames and material samples, preserving object scale and relative part transforms. The evaluator performs no scale fitting or independent part fitting. Reference and candidate captures use this alignment and shared cameras.

| Factor | Measurements |
| --- | --- |
| Shape | Bidirectional surface correspondence, surface F-score, precision, recall and distance statistics |
| Rigid parts | Corresponding body geometry, body count, relative arrangement and enabled/kinematic states |
| Joints | Joint count, type, connectivity, enabled state, anchors, relative frames, axes and limits |
| Materials | Supported surface properties, UV textures and authored friction, restitution, density and combine modes |

The structured `composite_scores` readout retains each composite's factors, earned credit, eligible denominators, measured counts, coverage and score bounds. Unknown measurements remain null and criteria absent from the reference are not applicable. Candidate omissions of eligible properties receive no credit. The [evaluation guide](https://github.com/hcltech-robotics/simreadybench/blob/main/docs/evaluation.md) defines the comparison bands and component calculations.

## Orientation-aligned comparison

![Reference eyeglasses and their reconstruction shown side by side through a shared camera orbit, using the verified whole-object alignment.](media/orientation-aligned-comparison.gif)

## Readouts and leaderboard

[Reconstruction readouts](https://huggingface.co/datasets/hcltech-robotics/simreadybench/tree/main/runs) group completed reconstructions by evaluation configuration. Each readout contains a browsable report, a score table and the generated assets and evaluation files:

```text
runs/<run-id>/
  report.html
  results.json
  scores.csv
  assets/<asset-id>/reconstruction-<n>/candidate/
  evidence/<asset-id>/reconstruction-<n>/
```

The dataset viewer displays matched reference and candidate images beside the score tables. The [leaderboard](https://huggingface.co/datasets/hcltech-robotics/simreadybench/blob/main/leaderboard.json) summarises complete NV25 evaluations with four reconstructions per asset, while individual readouts retain the measurements for every completed reconstruction.

## Evaluate a model

Use Python 3.12 in a Linux or WSL shell. Set `DAYTONA_API_KEY` and the credentials required by the model provider, then install the package and import the published runtime:

```sh
git clone https://github.com/hcltech-robotics/simreadybench.git
cd simreadybench
python3.12 -m venv .venv
. .venv/bin/activate
python -m pip install '.[daytona,publish]'
simready-bench setup daytona --output ../simreadybench-runtime.json
```

Run the benchmark with the saved runtime configuration:

```sh
simready-bench benchmark nvidia-25 --repetitions 4 \
  --model MODEL_ID --harness pi --deploy daytona \
  --runtime-images ../simreadybench-runtime.json \
  --authoring-turns 64 --workers 4 --supervise \
  --output-dir ../simreadybench-results
```

Replace `MODEL_ID` with the model identifier and add `--skills PATH` for an authoring skill condition. Add `--max-spend-usd AMOUNT` to set a separate monetary ceiling. For local execution, follow the deployment guide to configure a Linux Docker host, then use `--deploy local` with its `local-runtime-images.json`. Add `--hf-repo your-org/your-results --hf-run your-run-id` to publish completed reconstruction readouts to a Hugging Face dataset.

The [quickstart](https://github.com/hcltech-robotics/simreadybench/blob/main/docs/quickstart.md) covers model evaluation and result inspection, and the [deployment guide](https://github.com/hcltech-robotics/simreadybench/blob/main/docs/deployment.md) describes the local and Daytona runtimes.

## Citation

```bibtex
@software{simreadybench2026,
  title = {SimReadyBench: Evaluating models, harnesses and skills
           for simulation-ready asset generation},
  author = {von Csefalvay, Chris and Doma, Pranav and
            Unni Krishnan, Achyuthan and Foldi, Tamas},
  organization = {HCLTech},
  year = {2026},
  version = {1.0.0},
  url = {https://github.com/hcltech-robotics/simreadybench}
}
```