Add evaluation results for LiquidAI/d1-Omni-2B

Summary

This PR adds the benchmark results published on the model card for LiquidAI/d1-Omni-2B to .eval_results/, following the Hugging Face Hub evaluation-results specification.

Benchmarks Added

• Decision Index (public suite) — 49.0

Benchmarks Skipped (Not Registered on Hub)

These benchmarks are on the model card but have no registered eval.yaml on the Hub:

• DecisionBench (Hanno Labs) v1, Decision Bench (decisionbench.ai) v4, Fast Decisions, JevBench, ImajevBench, WorkflowEvals, Liquid's internal board
• The image benchmarks (AI2D, BLINK, MMBench, MME, ...)
• The audio tests (SLURP, Speech-MASSIVE, FLEURS, ESC-50, Belebele, decision suites read aloud)

They can be added once their authors register an eval.yaml on the Hub.

Mapping notes

• Typed Decisions (63.1): the card's metric and scale are unclear (the Hub task is accuracy on a 0-1 scale), so it's better to score it independently as was done for d1-3B.
• MMMU-Pro (24.1): scored as a decision over the options on at most 1,000 rows, which isn't the standard setup, and the variant isn't stated.

Source

• Model card: https://huggingface.co/LiquidAI/d1-Omni-2B
• Blog: https://www.liquid.ai/blog/open-d1

Files Added

• .eval_results/d1-Omni-2B.yaml

Verification

The scores come from the benchmark tables on the model card, as measured by Liquid AI. No verified token is included, because these were not run via HF Jobs with inspect-ai.

To upload:

hf upload LiquidAI/d1-Omni-2B .eval_results .eval_results --create-pr --commit-message "Add evaluation results"