Hugging Face's logo

marsnext
/
Phobos-1

Text Generation
PyTorch
Safetensors
English
phobos_cortical_moe
marsnext
phobos-1
conversational
code-generation
math-reasoning
system-2-reasoning
mixture-of-experts
cortical-moe
dynamic-mitosis
deepinfer-engram
multi-token-prediction
openai-compatible
edge-inference
Model card Files
xet
 Community
Phobos-1 / multi_model_head_to_head_benchmark.json
devMartianLabs's picture
devMartianLabs
Add head-to-head multi-model benchmark results and scripts (multi_model_head_to_head_benchmark.json)
1ff2bbc verified
Raw History Blame Contribute Delete
7.7 kB
{
"timestamp_utc": "2026-11-08T11:31:53Z",
"elapsed_sec": 10.31,
"leaderboard": [
{
"model_id": "marsnext/Phobos-1 (Final Post-Trained)",
"org": "MarsNext",
"total_params_m": 96.26,
"active_params_m": 52.22,
"train_tokens": "1.56B",
"mc20_acc_pct": 35.0,
"mc20_correct": "7/20",
"gen25_score_pct": 80.0,
"gen25_passed": "20/25",
"python_pass_at_1": "3/5 (60%)",
"sql_synthesis": "4/5 (80%)",
"stem_qa": "4/5 (80%)",
"math_symbolic": "4/5 (80%)",
"think_planning": "5/5 (100%)",
"combined_avg_pct": 57.5,
"speed_tok_per_s": 17.49
},
{
"model_id": "marsnext/Phobos-1 (Base Step 33,977)",
"org": "MarsNext",
"total_params_m": 96.26,
"active_params_m": 52.22,
"train_tokens": "1.56B",
"mc20_acc_pct": 35.0,
"mc20_correct": "7/20",
"gen25_score_pct": 68.0,
"gen25_passed": "17/25",
"python_pass_at_1": "3/5 (60%)",
"sql_synthesis": "4/5 (80%)",
"stem_qa": "3/5 (60%)",
"math_symbolic": "2/5 (40%)",
"think_planning": "5/5 (100%)",
"combined_avg_pct": 51.5,
"speed_tok_per_s": 18.52
},
{
"model_id": "EleutherAI/pythia-70m-deduped",
"org": "EleutherAI",
"total_params_m": 70.43,
"active_params_m": 70.43,
"train_tokens": "300B",
"mc20_acc_pct": 30.0,
"mc20_correct": "6/20",
"gen25_score_pct": 12.0,
"gen25_passed": "3/25",
"python_pass_at_1": "0/5 (0%)",
"sql_synthesis": "0/5 (0%)",
"stem_qa": "1/5 (20%)",
"math_symbolic": "1/5 (20%)",
"think_planning": "1/5 (20%)",
"combined_avg_pct": 21.0,
"speed_tok_per_s": 101.6
},
{
"model_id": "openai-community/gpt2",
"org": "OpenAI",
"total_params_m": 124.44,
"active_params_m": 124.44,
"train_tokens": "40B",
"mc20_acc_pct": 40.0,
"mc20_correct": "8/20",
"gen25_score_pct": 20.0,
"gen25_passed": "5/25",
"python_pass_at_1": "0/5 (0%)",
"sql_synthesis": "1/5 (20%)",
"stem_qa": "1/5 (20%)",
"math_symbolic": "1/5 (20%)",
"think_planning": "2/5 (40%)",
"combined_avg_pct": 30.0,
"speed_tok_per_s": 70.3
},
{
"model_id": "EleutherAI/pythia-160m-deduped",
"org": "EleutherAI",
"total_params_m": 162.32,
"active_params_m": 162.32,
"train_tokens": "300B",
"mc20_acc_pct": 25.0,
"mc20_correct": "5/20",
"gen25_score_pct": 12.0,
"gen25_passed": "3/25",
"python_pass_at_1": "0/5 (0%)",
"sql_synthesis": "0/5 (0%)",
"stem_qa": "1/5 (20%)",
"math_symbolic": "0/5 (0%)",
"think_planning": "2/5 (40%)",
"combined_avg_pct": 18.5,
"speed_tok_per_s": 67.1
},
{
"model_id": "HuggingFaceTB/SmolLM2-135M",
"org": "HuggingFace",
"total_params_m": 134.52,
"active_params_m": 134.52,
"train_tokens": "2,000B",
"mc20_acc_pct": 60.0,
"mc20_correct": "12/20",
"gen25_score_pct": 36.0,
"gen25_passed": "9/25",
"python_pass_at_1": "0/5 (0%)",
"sql_synthesis": "2/5 (40%)",
"stem_qa": "4/5 (80%)",
"math_symbolic": "1/5 (20%)",
"think_planning": "2/5 (40%)",
"combined_avg_pct": 48.0,
"speed_tok_per_s": 25.8
}
],
"phobos_mc_final": {
"correct": 7,
"total": 20,
"acc_pct": 35.0,
"details": [
{
"id": "ARC-01",
"pred": 0,
"ans": 1,
"correct": false
},
{
"id": "ARC-02",
"pred": 2,
"ans": 2,
"correct": true
},
{
"id": "ARC-03",
"pred": 0,
"ans": 0,
"correct": true
},
{
"id": "ARC-04",
"pred": 1,
"ans": 1,
"correct": true
},
{
"id": "ARC-05",
"pred": 1,
"ans": 2,
"correct": false
},
{
"id": "SCIQ-01",
"pred": 1,
"ans": 2,
"correct": false
},
{
"id": "SCIQ-02",
"pred": 1,
"ans": 1,
"correct": true
},
{
"id": "SCIQ-03",
"pred": 1,
"ans": 2,
"correct": false
},
{
"id": "SCIQ-04",
"pred": 2,
"ans": 1,
"correct": false
},
{
"id": "SCIQ-05",
"pred": 2,
"ans": 2,
"correct": true
},
{
"id": "PIQA-01",
"pred": 1,
"ans": 1,
"correct": true
},
{
"id": "PIQA-02",
"pred": 0,
"ans": 1,
"correct": false
},
{
"id": "PIQA-03",
"pred": 3,
"ans": 1,
"correct": false
},
{
"id": "CS-01",
"pred": 2,
"ans": 1,
"correct": false
},
{
"id": "CS-02",
"pred": 0,
"ans": 2,
"correct": false
},
{
"id": "CS-03",
"pred": 2,
"ans": 1,
"correct": false
},
{
"id": "CS-04",
"pred": 1,
"ans": 1,
"correct": true
},
{
"id": "MATH-MC1",
"pred": 0,
"ans": 1,
"correct": false
},
{
"id": "MATH-MC2",
"pred": 2,
"ans": 1,
"correct": false
},
{
"id": "MATH-MC3",
"pred": 3,
"ans": 0,
"correct": false
}
]
},
"phobos_mc_base": {
"correct": 7,
"total": 20,
"acc_pct": 35.0,
"details": [
{
"id": "ARC-01",
"pred": 0,
"ans": 1,
"correct": false
},
{
"id": "ARC-02",
"pred": 2,
"ans": 2,
"correct": true
},
{
"id": "ARC-03",
"pred": 0,
"ans": 0,
"correct": true
},
{
"id": "ARC-04",
"pred": 1,
"ans": 1,
"correct": true
},
{
"id": "ARC-05",
"pred": 1,
"ans": 2,
"correct": false
},
{
"id": "SCIQ-01",
"pred": 1,
"ans": 2,
"correct": false
},
{
"id": "SCIQ-02",
"pred": 1,
"ans": 1,
"correct": true
},
{
"id": "SCIQ-03",
"pred": 1,
"ans": 2,
"correct": false
},
{
"id": "SCIQ-04",
"pred": 2,
"ans": 1,
"correct": false
},
{
"id": "SCIQ-05",
"pred": 2,
"ans": 2,
"correct": true
},
{
"id": "PIQA-01",
"pred": 1,
"ans": 1,
"correct": true
},
{
"id": "PIQA-02",
"pred": 0,
"ans": 1,
"correct": false
},
{
"id": "PIQA-03",
"pred": 3,
"ans": 1,
"correct": false
},
{
"id": "CS-01",
"pred": 2,
"ans": 1,
"correct": false
},
{
"id": "CS-02",
"pred": 0,
"ans": 2,
"correct": false
},
{
"id": "CS-03",
"pred": 2,
"ans": 1,
"correct": false
},
{
"id": "CS-04",
"pred": 1,
"ans": 1,
"correct": true
},
{
"id": "MATH-MC1",
"pred": 2,
"ans": 1,
"correct": false
},
{
"id": "MATH-MC2",
"pred": 2,
"ans": 1,
"correct": false
},
{
"id": "MATH-MC3",
"pred": 3,
"ans": 0,
"correct": false
}
]
}
}