🏆 HF Hub Benchmark Dashboard
Last updated: 2026-06-01T15:26:36 UTC · auto-refreshes every 6h
29 benchmarks
296 models
593 entries
25 active
4 empty
🧠 Knowledge · 3 benchmarks · 88 models
GPQA
50 entries↗ View on Hub
| Model | Score | In $/1M | Out $/1M | Context | TTFT | Throughput | License | Params | Providers | |
|---|---|---|---|---|---|---|---|---|---|---|
| 🥇 | moonshotai/Kimi-K2.6 | 90.5 | $0.75 | $3.40 | 262.1K | 422 ms | 81 t/s | other | 1059B | novitatogetherfireworks-ai +2 |
| 🥈 | deepseek-ai/DeepSeek-V4-Pro | 90.1 | $1.60 | $3.38 | 1.0M | 491 ms | 42 t/s | mit | 862B | novitatogetherfireworks-ai +2 |
| 🥉 | FINAL-Bench/Darwin-28B-REASON | 89.39 | — | — | — | — | — | apache-2.0 | 27B | — |
| 4 | OrionLLM/GRM-2.6-Opus | 89.2 | — | — | — | — | — | apache-2.0 | 28B | — |
| 5 | FINAL-Bench/Darwin-28B-Opus | 88.89 | — | — | — | — | — | apache-2.0 | 28B | — |
| 6 | Qwen/Qwen3.5-397B-A17B | 88.4 | $0.49 | $3.60 | 262.1K | 385 ms | 88 t/s | apache-2.0 | 403B | novitatogetherfeatherless-ai +3 |
| 7 | FINAL-Bench/Darwin-36B-Opus | 88.4 | — | — | — | — | — | apache-2.0 | 35B | — |
| 8 | FINAL-Bench/Darwin-60B-DUO | 88.38 | — | — | — | — | — | gemma | — | — |
| 9 | OrionLLM/GRM-2.6-Plus | 88.3 | — | — | — | — | — | apache-2.0 | 28B | — |
| 10 | inclusionAI/Ring-2.6-1T | 88.27 | — | — | — | — | — | mit | 1026B | — |
| 11 | deepseek-ai/DeepSeek-V4-Flash | 88.1 | $0.14 | $0.28 | 1.0M | 563 ms | 110 t/s | mit | 158B | novitafireworks-aifeatherless-ai +1 |
| 12 | Qwen/Qwen3.6-27B | 87.8 | — | — | — | — | — | apache-2.0 | 28B | — |
| 13 | moonshotai/Kimi-K2.5 | 87.6 | $0.60 | $3.00 | 262.1K | 906 ms | 35 t/s | other | 1059B | novitafireworks-aifeatherless-ai |
| 14 | tencent/Hy3-preview | 87.2 | — | — | — | — | — | other | 299B | — |
| 15 | FINAL-Bench/Darwin-27B-Opus | 86.9 | — | — | — | — | — | apache-2.0 | 28B | — |
| 16 | Qwen/Qwen3.5-122B-A10B | 86.6 | $0.29 | $2.40 | 262.1K | 281 ms | 90 t/s | apache-2.0 | 125B | novitadeepinfra |
| 17 | zai-org/GLM-5.1 | 86.2 | $1.05 | $3.50 | 202.8K | 940 ms | 34 t/s | mit | 754B | togetherfireworks-aifeatherless-ai +2 |
| 18 | zai-org/GLM-5 | 86 | $1.00 | $3.20 | 202.8K | 575 ms | 88 t/s | mit | 754B | novitatogetherfeatherless-ai +1 |
| 19 | Qwen/Qwen3.6-35B-A3B | 86 | $0.15 | $0.95 | 262.1K | 186 ms | 146 t/s | apache-2.0 | 36B | featherless-aideepinfra |
| 20 | FINAL-Bench/Darwin-31B-Opus | 85.9 | — | — | — | — | — | apache-2.0 | 33B | — |
| 21 | zai-org/GLM-4.7 | 85.7 | $0.60 | $2.20 | 204.8K | 132 ms | 414 t/s | mit | 358B | novitacerebrasfeatherless-ai +1 |
| 22 | Qwen/Qwen3.5-27B | 85.5 | $0.30 | $2.40 | 262.1K | 892 ms | 47 t/s | apache-2.0 | 28B | novitafeatherless-ai |
| 23 | MiniMaxAI/MiniMax-M2.5 | 85.2 | $0.30 | $1.20 | 204.8K | 751 ms | 105 t/s | other | 229B | novitafireworks-aifeatherless-ai |
| 24 | moonshotai/Kimi-K2-Thinking | 84.5 | $0.60 | $2.50 | 262.1K | 1,111 ms | 49 t/s | other | 1058B | novitafeatherless-ai |
| 25 | FINAL-Bench/Darwin-9B-NEG | 84.34 | — | — | — | — | — | apache-2.0 | 10B | — |
| 26 | google/gemma-4-31B-it | 84.3 | $0.13 | $0.38 | 262.1K | 402 ms | 79 t/s | apache-2.0 | 33B | novitatogetherfeatherless-ai +1 |
| 27 | Qwen/Qwen3.5-35B-A3B | 84.2 | $0.25 | $2.00 | 262.1K | 697 ms | 96 t/s | apache-2.0 | 36B | novita |
| 28 | Nanbeige/Nanbeige4.1-3B | 83.8 | — | — | — | — | — | apache-2.0 | 4B | — |
| 29 | stepfun-ai/Step-3.5-Flash | 83.5 | $0.10 | $0.30 | 262.1K | 277 ms | 43 t/s | apache-2.0 | 199B | featherless-aideepinfra |
| 30 | nvidia/NVIDIA-Nemotron-3-Super-120B-A12B-BF16 | 82.7 | — | — | — | — | — | other | 124B | — |
| 31 | RedHatAI/NVIDIA-Nemotron-3-Super-120B-A12B-BF16 | 82.7 | — | — | — | — | — | other | 124B | — |
| 32 | deepseek-ai/DeepSeek-V3.2 | 82.4 | $0.27 | $0.40 | 163.8K | 1,853 ms | 33 t/s | mit | 685B | novitafeatherless-ai |
| 33 | google/gemma-4-26B-A4B-it | 82.3 | $0.07 | $0.34 | 262.1K | 307 ms | 40 t/s | apache-2.0 | 27B | novitafeatherless-aideepinfra |
| 34 | Qwen/Qwen3.5-9B | 81.7 | $0.12 | $0.18 | 262.1K | 236 ms | 84 t/s | apache-2.0 | 10B | togetherfeatherless-aiovhcloud |
| 35 | openai/gpt-oss-120b | 80.9 | $0.05 | $0.25 | 131.1K | 165 ms | 1120 t/s | apache-2.0 | 120B | groqnovitacerebras +7 |
| 36 | meituan-longcat/LongCat-Flash-Thinking-2601 | 80.5 | — | — | — | — | — | mit | 562B | — |
| 37 | LGAI-EXAONE/EXAONE-4.5-33B | 80.5 | — | — | — | — | — | other | 34B | — |
| 38 | openai/gpt-oss-120b | 80.1 | $0.05 | $0.25 | 131.1K | 165 ms | 1120 t/s | apache-2.0 | 120B | groqnovitacerebras +7 |
| 39 | nvidia/NVIDIA-Nemotron-3-Super-120B-A12B-BF16 | 79.23 | — | — | — | — | — | other | 124B | — |
| 40 | RedHatAI/NVIDIA-Nemotron-3-Super-120B-A12B-BF16 | 79.23 | — | — | — | — | — | other | 124B | — |
| 41 | LGAI-EXAONE/K-EXAONE-236B-A23B | 79.1 | — | — | — | — | — | other | 237B | — |
| 42 | OrionLLM/GRM-2.5 | 76.7 | — | — | — | — | — | apache-2.0 | 5B | — |
| 43 | arcee-ai/Trinity-Large-Thinking | 76.3 | — | — | — | — | — | other | 399B | — |
| 44 | Qwen/Qwen3.5-4B | 76.2 | — | — | — | — | — | apache-2.0 | 5B | — |
| 45 | nvidia/Nemotron-Cascade-2-30B-A3B | 76.1 | — | — | — | — | — | other | 32B | — |
| 46 | zai-org/GLM-4.7-Flash | 75.2 | — | — | — | 2,324 ms | 61 t/s | mit | 31B | featherless-aizai-org |
| 47 | jdopensource/JoyAI-LLM-Flash | 74.43 | — | — | — | — | — | — | 49B | — |
| 48 | openai/gpt-oss-20b | 74.2 | $0.04 | $0.15 | 131.1K | 241 ms | 680 t/s | apache-2.0 | 22B | groqnovitanscale +4 |
| 49 | openai/gpt-oss-120b | 73.5 | $0.05 | $0.25 | 131.1K | 165 ms | 1120 t/s | apache-2.0 | 120B | groqnovitacerebras +7 |
| 50 | openai/gpt-oss-120b | 73.1 | $0.05 | $0.25 | 131.1K | 165 ms | 1120 t/s | apache-2.0 | 120B | groqnovitacerebras +7 |