BenchmarkActivecategory · reasoning
quality57
Updated 9 h ago · first seen 11 Sept 2026
bench_01M293SPE8YX533DB8C6AE1Q1S
- Metric
- accuracy · %
- Direction
- Higher is better
- Results
- 614
- Leader
- gpt-6-astra-xhigh 96.26%
Score history · Step 3.5 Flash 1 row
Not enough history to chart — a single observation (83.13% on 11 Sept 2026). Rows under different configurations count separately; the list below shows each one.
- 83.13%aa_slug=step-3-5-flash-0202 · variant=GPQA Diamond · evaluator=Artificial Analysis · index_version=4.311 Sept 2026
Leaderboard 614 current results
Select models with +, then open Compare.
| # | Model | Score | Config | Evaluated | Source | Actions |
|---|---|---|---|---|---|---|
| #501gemma-4-E2BGoogle | 43.33% | aa_slug=gemma-4-e2b · variant=GPQA Diamond · evaluator=Artificial Analysis · index_version=4.3 | — (obs. 11 Sept 2026) | artificialanalysis.aiT2 | History | |
| #502nova-liteAmazon Web Services | 43.33% | aa_slug=nova-lite · variant=GPQA Diamond · evaluator=Artificial Analysis · index_version=4.3 | — (obs. 11 Sept 2026) | artificialanalysis.aiT2 | History | |
| #503llama-3-2-instruct-90b-visionMeta AI | 43.23% | aa_slug=llama-3-2-instruct-90b-vision · variant=GPQA Diamond · evaluator=Artificial Analysis · index_version=4.3 | — (obs. 11 Sept 2026) | artificialanalysis.aiT2 | History | |
| #504Gemma 3 27BGoogle | 42.83% | aa_slug=gemma-3-27b · variant=GPQA Diamond · evaluator=Artificial Analysis · index_version=4.3 | — (obs. 11 Sept 2026) | artificialanalysis.aiT2 | History | |
| #505gpt-5-nano-minimalOpenAI | 42.83% | aa_slug=gpt-5-nano-minimal · variant=GPQA Diamond · evaluator=Artificial Analysis · index_version=4.3 | — (obs. 11 Sept 2026) | artificialanalysis.aiT2 | History | |
| #506Qwen3 VL 8B InstructQwen | 42.73% | aa_slug=qwen3-vl-8b-instruct · variant=GPQA Diamond · evaluator=Artificial Analysis · index_version=4.3 | — (obs. 11 Sept 2026) | artificialanalysis.aiT2 | History | |
| #507jamba-1-5-largeAI21 Labs | 42.73% | aa_slug=jamba-1-5-large · variant=GPQA Diamond · evaluator=Artificial Analysis · index_version=4.3 | — (obs. 11 Sept 2026) | artificialanalysis.aiT2 | History | |
| #508gpt-4o-miniOpenAI | 42.63% | aa_slug=gpt-4o-mini · variant=GPQA Diamond · evaluator=Artificial Analysis · index_version=4.3 | — (obs. 11 Sept 2026) | artificialanalysis.aiT2 | History | |
| #509Molmo2-8BAllen Institute for AI | 42.53% | aa_slug=molmo2-8b · variant=GPQA Diamond · evaluator=Artificial Analysis · index_version=4.3 | — (obs. 11 Sept 2026) | artificialanalysis.aiT2 | History | |
| #510exaone-4-0-1-2bLG AI Research | 42.42% | aa_slug=exaone-4-0-1-2b · variant=GPQA Diamond · evaluator=Artificial Analysis · index_version=4.3 | — (obs. 11 Sept 2026) | artificialanalysis.aiT2 | History | |
| #511Mistral SabaMistral AI | 42.42% | aa_slug=mistral-saba · variant=GPQA Diamond · evaluator=Artificial Analysis · index_version=4.3 | — (obs. 11 Sept 2026) | artificialanalysis.aiT2 | History | |
| #512deepseek-v2-5DeepSeek | 42.32% | aa_slug=deepseek-v2-5 · variant=GPQA Diamond · evaluator=Artificial Analysis · index_version=4.3 | — (obs. 11 Sept 2026) | artificialanalysis.aiT2 | History | |
| #513Qwen2.5 Coder 32B InstructQwen | 41.72% | aa_slug=qwen2-5-coder-32b-instruct · variant=GPQA Diamond · evaluator=Artificial Analysis · index_version=4.3 | — (obs. 11 Sept 2026) | artificialanalysis.aiT2 | History | |
| #514sarvam-m-reasoningSarvam | 41.62% | aa_slug=sarvam-m-reasoning · variant=GPQA Diamond · evaluator=Artificial Analysis · index_version=4.3 | — (obs. 11 Sept 2026) | artificialanalysis.aiT2 | History | |
| #515granite-4-0-h-smallIBM | 41.62% | aa_slug=granite-4-0-h-small · variant=GPQA Diamond · evaluator=Artificial Analysis · index_version=4.3 | — (obs. 11 Sept 2026) | artificialanalysis.aiT2 | History | |
| #516devstral-smallMistral AI | 41.41% | aa_slug=devstral-small · variant=GPQA Diamond · evaluator=Artificial Analysis · index_version=4.3 | — (obs. 11 Sept 2026) | artificialanalysis.aiT2 | History | |
| #517Kimi-Linear-48B-A3B-InstructMoonshot AI | 41.21% | aa_slug=kimi-linear-48b-a3b-instruct · variant=GPQA Diamond · evaluator=Artificial Analysis · index_version=4.3 | — (obs. 11 Sept 2026) | artificialanalysis.aiT2 | History | |
| #518qwen-turboAlibaba Group | 41.01% | aa_slug=qwen-turbo · variant=GPQA Diamond · evaluator=Artificial Analysis · index_version=4.3 | — (obs. 11 Sept 2026) | artificialanalysis.aiT2 | History | |
| #519Llama-3.1-70BMeta AI | 40.91% | aa_slug=llama-3-1-instruct-70b · variant=GPQA Diamond · evaluator=Artificial Analysis · index_version=4.3 | — (obs. 11 Sept 2026) | artificialanalysis.aiT2 | History | |
| #520Claude Haiku 3.5Anthropic | 40.81% | aa_slug=claude-3-5-haiku · variant=GPQA Diamond · evaluator=Artificial Analysis · index_version=4.3 | — (obs. 11 Sept 2026) | artificialanalysis.aiT2 | History | |
| #521llama-3-1-nemotron-nano-4b-reasoningNVIDIA | 40.81% | aa_slug=llama-3-1-nemotron-nano-4b-reasoning · variant=GPQA Diamond · evaluator=Artificial Analysis · index_version=4.3 | — (obs. 11 Sept 2026) | artificialanalysis.aiT2 | History | |
| #522gemma-4-e2b-non-reasoningGoogle | 40.51% | aa_slug=gemma-4-e2b-non-reasoning · variant=GPQA Diamond · evaluator=Artificial Analysis · index_version=4.3 | — (obs. 11 Sept 2026) | artificialanalysis.aiT2 | History | |
| #523R1 Distill Llama 70BDeepSeek | 40.2% | aa_slug=deepseek-r1-distill-llama-70b · variant=GPQA Diamond · evaluator=Artificial Analysis · index_version=4.3 | — (obs. 11 Sept 2026) | artificialanalysis.aiT2 | History | |
| #524Hermes 3 70B InstructNous Research | 40.1% | aa_slug=hermes-3-llama-3-1-70b · variant=GPQA Diamond · evaluator=Artificial Analysis · index_version=4.3 | — (obs. 11 Sept 2026) | artificialanalysis.aiT2 | History | |
| #525Olmo-3-7B-InstructAllen Institute for AI | 40% | aa_slug=olmo-3-7b-instruct · variant=GPQA Diamond · evaluator=Artificial Analysis · index_version=4.3 | — (obs. 11 Sept 2026) | artificialanalysis.aiT2 | History | |
| #526claude-3-sonnetAnthropic | 40% | aa_slug=claude-3-sonnet · variant=GPQA Diamond · evaluator=Artificial Analysis · index_version=4.3 | — (obs. 11 Sept 2026) | artificialanalysis.aiT2 | History | |
| #527Nemotron 3 Nano 30B A3BNVIDIA | 39.9% | aa_slug=nvidia-nemotron-3-nano-30b-a3b · variant=GPQA Diamond · evaluator=Artificial Analysis · index_version=4.3 | — (obs. 11 Sept 2026) | artificialanalysis.aiT2 | History | |
| #528qwen3-4b-instructAlibaba Group | 39.8% | aa_slug=qwen3-4b-instruct · variant=GPQA Diamond · evaluator=Artificial Analysis · index_version=4.3 | — (obs. 11 Sept 2026) | artificialanalysis.aiT2 | History | |
| #529jamba-1-7-largeAI21 Labs | 38.99% | aa_slug=jamba-1-7-large · variant=GPQA Diamond · evaluator=Artificial Analysis · index_version=4.3 | — (obs. 11 Sept 2026) | artificialanalysis.aiT2 | History | |
| #530jamba-1-6-largeAI21 Labs | 38.69% | aa_slug=jamba-1-6-large · variant=GPQA Diamond · evaluator=Artificial Analysis · index_version=4.3 | — (obs. 11 Sept 2026) | artificialanalysis.aiT2 | History | |
| #531deephermes-3-mistral-24b-previewNous Research | 38.18% | aa_slug=deephermes-3-mistral-24b-preview · variant=GPQA Diamond · evaluator=Artificial Analysis · index_version=4.3 | — (obs. 11 Sept 2026) | artificialanalysis.aiT2 | History | |
| #532mistral-smallMistral AI | 38.08% | aa_slug=mistral-small · variant=GPQA Diamond · evaluator=Artificial Analysis · index_version=4.3 | — (obs. 11 Sept 2026) | artificialanalysis.aiT2 | History | |
| #533llama-3-instruct-70bMeta AI | 37.88% | aa_slug=llama-3-instruct-70b · variant=GPQA Diamond · evaluator=Artificial Analysis · index_version=4.3 | — (obs. 11 Sept 2026) | artificialanalysis.aiT2 | History | |
| #534Claude 3 HaikuAnthropic | 37.37% | aa_slug=claude-3-haiku · variant=GPQA Diamond · evaluator=Artificial Analysis · index_version=4.3 | — (obs. 11 Sept 2026) | artificialanalysis.aiT2 | History | |
| #535qwen2-72b-instructAlibaba Group | 37.07% | aa_slug=qwen2-72b-instruct · variant=GPQA Diamond · evaluator=Artificial Analysis · index_version=4.3 | — (obs. 11 Sept 2026) | artificialanalysis.aiT2 | History | |
| #536Qwen3-VL-4B-InstructQwen | 37.07% | aa_slug=qwen3-vl-4b-instruct · variant=GPQA Diamond · evaluator=Artificial Analysis · index_version=4.3 | — (obs. 11 Sept 2026) | artificialanalysis.aiT2 | History | |
| #537gemini-1-5-pro-may-2024Google | 37.07% | aa_slug=gemini-1-5-pro-may-2024 · variant=GPQA Diamond · evaluator=Artificial Analysis · index_version=4.3 | — (obs. 11 Sept 2026) | artificialanalysis.aiT2 | History | |
| #538gemini-1-5-flash-8bGoogle | 35.86% | aa_slug=gemini-1-5-flash-8b · variant=GPQA Diamond · evaluator=Artificial Analysis · index_version=4.3 | — (obs. 11 Sept 2026) | artificialanalysis.aiT2 | History | |
| #539nova-microAmazon Web Services | 35.76% | aa_slug=nova-micro · variant=GPQA Diamond · evaluator=Artificial Analysis · index_version=4.3 | — (obs. 11 Sept 2026) | artificialanalysis.aiT2 | History | |
| #540Ministral 3 3BMistral AI | 35.76% | aa_slug=ministral-3-3b · variant=GPQA Diamond · evaluator=Artificial Analysis · index_version=4.3 | — (obs. 11 Sept 2026) | artificialanalysis.aiT2 | History | |
| #541qwen3-1.7b-instruct-reasoningAlibaba Group | 35.56% | aa_slug=qwen3-1.7b-instruct-reasoning · variant=GPQA Diamond · evaluator=Artificial Analysis · index_version=4.3 | — (obs. 11 Sept 2026) | artificialanalysis.aiT2 | History | |
| #542Mistral LargeMistral AI | 35.05% | aa_slug=mistral-large · variant=GPQA Diamond · evaluator=Artificial Analysis · index_version=4.3 | — (obs. 11 Sept 2026) | artificialanalysis.aiT2 | History | |
| #543Gemma 3 12BGoogle | 34.95% | aa_slug=gemma-3-12b · variant=GPQA Diamond · evaluator=Artificial Analysis · index_version=4.3 | — (obs. 11 Sept 2026) | artificialanalysis.aiT2 | History | |
| #544mistral-mediumMistral AI | 34.95% | aa_slug=mistral-medium · variant=GPQA Diamond · evaluator=Artificial Analysis · index_version=4.3 | — (obs. 11 Sept 2026) | artificialanalysis.aiT2 | History | |
| #545gpt-4OpenAI | 34.95% | aa_slug=gpt-4 · variant=GPQA Diamond · evaluator=Artificial Analysis · index_version=4.3 | — (obs. 11 Sept 2026) | artificialanalysis.aiT2 | History | |
| #546lfm2-8b-a1bLiquid AI | 34.44% | aa_slug=lfm2-8b-a1b · variant=GPQA Diamond · evaluator=Artificial Analysis · index_version=4.3 | — (obs. 11 Sept 2026) | artificialanalysis.aiT2 | History | |
| #547claude-2Anthropic | 34.44% | aa_slug=claude-2 · variant=GPQA Diamond · evaluator=Artificial Analysis · index_version=4.3 | — (obs. 11 Sept 2026) | artificialanalysis.aiT2 | History | |
| #548lfm2-5-1-2b-thinkingLiquid AI | 33.94% | aa_slug=lfm2-5-1-2b-thinking · variant=GPQA Diamond · evaluator=Artificial Analysis · index_version=4.3 | — (obs. 11 Sept 2026) | artificialanalysis.aiT2 | History | |
| #549Qwen2.5-Coder-7BQwen | 33.94% | aa_slug=qwen2-5-coder-7b-instruct · variant=GPQA Diamond · evaluator=Artificial Analysis · index_version=4.3 | — (obs. 11 Sept 2026) | artificialanalysis.aiT2 | History | |
| #550granite-3-3-8b-instructIBM | 33.84% | aa_slug=granite-3-3-8b-instruct · variant=GPQA Diamond · evaluator=Artificial Analysis · index_version=4.3 | — (obs. 11 Sept 2026) | artificialanalysis.aiT2 | History | |
| #551Granite 4.0 MicroIBM | 33.64% | aa_slug=granite-4-0-micro · variant=GPQA Diamond · evaluator=Artificial Analysis · index_version=4.3 | — (obs. 11 Sept 2026) | artificialanalysis.aiT2 | History | |
| #552jamba-reasoning-3bAI21 Labs | 33.33% | aa_slug=jamba-reasoning-3b · variant=GPQA Diamond · evaluator=Artificial Analysis · index_version=4.3 | — (obs. 11 Sept 2026) | artificialanalysis.aiT2 | History | |
| #553Mixtral 8x22BMistral AI | 33.23% | aa_slug=mistral-8x22b-instruct · variant=GPQA Diamond · evaluator=Artificial Analysis · index_version=4.3 | — (obs. 11 Sept 2026) | artificialanalysis.aiT2 | History | |
| #554phi-4-miniMicrosoft | 33.13% | aa_slug=phi-4-mini · variant=GPQA Diamond · evaluator=Artificial Analysis · index_version=4.3 | — (obs. 11 Sept 2026) | artificialanalysis.aiT2 | History | |
| #555dbrxDatabricks | 33.13% | aa_slug=dbrx · variant=GPQA Diamond · evaluator=Artificial Analysis · index_version=4.3 | — (obs. 11 Sept 2026) | artificialanalysis.aiT2 | History | |
| #556claude-instantAnthropic | 33.03% | aa_slug=claude-instant · variant=GPQA Diamond · evaluator=Artificial Analysis · index_version=4.3 | — (obs. 11 Sept 2026) | artificialanalysis.aiT2 | History | |
| #557olmo-2-32bAllen Institute for AI | 32.83% | aa_slug=olmo-2-32b · variant=GPQA Diamond · evaluator=Artificial Analysis · index_version=4.3 | — (obs. 11 Sept 2026) | artificialanalysis.aiT2 | History | |
| #558lfm-40bLiquid AI | 32.73% | aa_slug=lfm-40b · variant=GPQA Diamond · evaluator=Artificial Analysis · index_version=4.3 | — (obs. 11 Sept 2026) | artificialanalysis.aiT2 | History | |
| #559llama-2-chat-70bMeta AI | 32.73% | aa_slug=llama-2-chat-70b · variant=GPQA Diamond · evaluator=Artificial Analysis · index_version=4.3 | — (obs. 11 Sept 2026) | artificialanalysis.aiT2 | History | |
| #560LFM2.5-1.2B-InstructLiquid AI | 32.63% | aa_slug=lfm2-5-1-2b-instruct · variant=GPQA Diamond · evaluator=Artificial Analysis · index_version=4.3 | — (obs. 11 Sept 2026) | artificialanalysis.aiT2 | History | |
| #561gemini-1-5-flash-may-2024Google | 32.42% | aa_slug=gemini-1-5-flash-may-2024 · variant=GPQA Diamond · evaluator=Artificial Analysis · index_version=4.3 | — (obs. 11 Sept 2026) | artificialanalysis.aiT2 | History | |
| #562command-r-plus-04-2024Cohere | 32.32% | aa_slug=command-r-plus-04-2024 · variant=GPQA Diamond · evaluator=Artificial Analysis · index_version=4.3 | — (obs. 11 Sept 2026) | artificialanalysis.aiT2 | History | |
| #563jamba-1-7-miniAI21 Labs | 32.22% | aa_slug=jamba-1-7-mini · variant=GPQA Diamond · evaluator=Artificial Analysis · index_version=4.3 | — (obs. 11 Sept 2026) | artificialanalysis.aiT2 | History | |
| #564llama-2-chat-13bMeta AI | 32.12% | aa_slug=llama-2-chat-13b · variant=GPQA Diamond · evaluator=Artificial Analysis · index_version=4.3 | — (obs. 11 Sept 2026) | artificialanalysis.aiT2 | History | |
| #565claude-21Anthropic | 31.92% | aa_slug=claude-21 · variant=GPQA Diamond · evaluator=Artificial Analysis · index_version=4.3 | — (obs. 11 Sept 2026) | artificialanalysis.aiT2 | History | |
| #566deepseek-coder-v2-liteDeepSeek | 31.92% | aa_slug=deepseek-coder-v2-lite · variant=GPQA Diamond · evaluator=Artificial Analysis · index_version=4.3 | — (obs. 11 Sept 2026) | artificialanalysis.aiT2 | History | |
| #567phi-3-miniMicrosoft | 31.92% | aa_slug=phi-3-mini · variant=GPQA Diamond · evaluator=Artificial Analysis · index_version=4.3 | — (obs. 11 Sept 2026) | artificialanalysis.aiT2 | History | |
| #568phi-4-multimodalMicrosoft | 31.52% | aa_slug=phi-4-multimodal · variant=GPQA Diamond · evaluator=Artificial Analysis · index_version=4.3 | — (obs. 11 Sept 2026) | artificialanalysis.aiT2 | History | |
| #569granite-4.1-3bIBM | 31.41% | aa_slug=granite-4-1-3b · variant=GPQA Diamond · evaluator=Artificial Analysis · index_version=4.3 | — (obs. 11 Sept 2026) | artificialanalysis.aiT2 | History | |
| #570lfm2-2-6bLiquid AI | 30.61% | aa_slug=lfm2-2-6b · variant=GPQA Diamond · evaluator=Artificial Analysis · index_version=4.3 | — (obs. 11 Sept 2026) | artificialanalysis.aiT2 | History | |
| #571tiny-aya-globalCohere | 30.51% | aa_slug=tiny-aya-global · variant=GPQA Diamond · evaluator=Artificial Analysis · index_version=4.3 | — (obs. 11 Sept 2026) | artificialanalysis.aiT2 | History | |
| #572minicpm-v4-6-1-3bOpenBMB | 30.51% | aa_slug=minicpm-v4-6-1-3b · variant=GPQA Diamond · evaluator=Artificial Analysis · index_version=4.3 | — (obs. 11 Sept 2026) | artificialanalysis.aiT2 | History | |
| #573jamba-1-5-miniAI21 Labs | 30.2% | aa_slug=jamba-1-5-mini · variant=GPQA Diamond · evaluator=Artificial Analysis · index_version=4.3 | — (obs. 11 Sept 2026) | artificialanalysis.aiT2 | History | |
| #574Mistral Small 1.0Mistral AI | 30.2% | aa_slug=mistral-small-2402 · variant=GPQA Diamond · evaluator=Artificial Analysis · index_version=4.3 | — (obs. 11 Sept 2026) | artificialanalysis.aiT2 | History | |
| #575deepseek-r1-distill-llama-8bDeepSeek | 30.2% | aa_slug=deepseek-r1-distill-llama-8b · variant=GPQA Diamond · evaluator=Artificial Analysis · index_version=4.3 | — (obs. 11 Sept 2026) | artificialanalysis.aiT2 | History | |
| #576jamba-1-6-miniAI21 Labs | 30% | aa_slug=jamba-1-6-mini · variant=GPQA Diamond · evaluator=Artificial Analysis · index_version=4.3 | — (obs. 11 Sept 2026) | artificialanalysis.aiT2 | History | |
| #577GPT-3.5 TurboOpenAI | 29.7% | aa_slug=gpt-35-turbo · variant=GPQA Diamond · evaluator=Artificial Analysis · index_version=4.3 | — (obs. 11 Sept 2026) | artificialanalysis.aiT2 | History | |
| #578gemma-3n-e4bGoogle | 29.6% | aa_slug=gemma-3n-e4b · variant=GPQA Diamond · evaluator=Artificial Analysis · index_version=4.3 | — (obs. 11 Sept 2026) | artificialanalysis.aiT2 | History | |
| #579llama-3-instruct-8bMeta AI | 29.6% | aa_slug=llama-3-instruct-8b · variant=GPQA Diamond · evaluator=Artificial Analysis · index_version=4.3 | — (obs. 11 Sept 2026) | artificialanalysis.aiT2 | History | |
| #580Mixtral 8x7BMistral AI | 29.19% | aa_slug=mixtral-8x7b-instruct · variant=GPQA Diamond · evaluator=Artificial Analysis · index_version=4.3 | — (obs. 11 Sept 2026) | artificialanalysis.aiT2 | History | |
| #581Gemma 3 4BGoogle | 29.09% | aa_slug=gemma-3-4b · variant=GPQA Diamond · evaluator=Artificial Analysis · index_version=4.3 | — (obs. 11 Sept 2026) | artificialanalysis.aiT2 | History | |
| #582qwen1.5-110b-chatAlibaba Group | 28.89% | aa_slug=qwen1.5-110b-chat · variant=GPQA Diamond · evaluator=Artificial Analysis · index_version=4.3 | — (obs. 11 Sept 2026) | artificialanalysis.aiT2 | History | |
| #583LFM2.5-VL-1.6BLiquid AI | 28.89% | aa_slug=lfm2-5-vl-1-6b · variant=GPQA Diamond · evaluator=Artificial Analysis · index_version=4.3 | — (obs. 11 Sept 2026) | artificialanalysis.aiT2 | History | |
| #584olmo-2-7bAllen Institute for AI | 28.79% | aa_slug=olmo-2-7b · variant=GPQA Diamond · evaluator=Artificial Analysis · index_version=4.3 | — (obs. 11 Sept 2026) | artificialanalysis.aiT2 | History | |
| #585command-r-03-2024Cohere | 28.38% | aa_slug=command-r-03-2024 · variant=GPQA Diamond · evaluator=Artificial Analysis · index_version=4.3 | — (obs. 11 Sept 2026) | artificialanalysis.aiT2 | History | |
| #586qwen3-1.7b-instructAlibaba Group | 28.28% | aa_slug=qwen3-1.7b-instruct · variant=GPQA Diamond · evaluator=Artificial Analysis · index_version=4.3 | — (obs. 11 Sept 2026) | artificialanalysis.aiT2 | History | |
| #587granite-4-0-nano-1bIBM | 28.08% | aa_slug=granite-4-0-nano-1b · variant=GPQA Diamond · evaluator=Artificial Analysis · index_version=4.3 | — (obs. 11 Sept 2026) | artificialanalysis.aiT2 | History | |
| #588minicpm5-1bOpenBMB | 27.78% | aa_slug=minicpm5-1b · variant=GPQA Diamond · evaluator=Artificial Analysis · index_version=4.3 | — (obs. 11 Sept 2026) | artificialanalysis.aiT2 | History | |
| #589gemma-3n-e4b-preview-0520Google | 27.78% | aa_slug=gemma-3n-e4b-preview-0520 · variant=GPQA Diamond · evaluator=Artificial Analysis · index_version=4.3 | — (obs. 11 Sept 2026) | artificialanalysis.aiT2 | History | |
| #590gemini-1-0-proGoogle | 27.68% | aa_slug=gemini-1-0-pro · variant=GPQA Diamond · evaluator=Artificial Analysis · index_version=4.3 | — (obs. 11 Sept 2026) | artificialanalysis.aiT2 | History | |
| #591apertus-70b-instructSwiss AI Initiative | 27.17% | aa_slug=apertus-70b-instruct · variant=GPQA Diamond · evaluator=Artificial Analysis · index_version=4.3 | — (obs. 11 Sept 2026) | artificialanalysis.aiT2 | History | |
| #592deephermes-3-llama-3-1-8b-previewNous Research | 26.97% | aa_slug=deephermes-3-llama-3-1-8b-preview · variant=GPQA Diamond · evaluator=Artificial Analysis · index_version=4.3 | — (obs. 11 Sept 2026) | artificialanalysis.aiT2 | History | |
| #593minicpm5-1b-non-reasoningOpenBMB | 26.87% | aa_slug=minicpm5-1b-non-reasoning · variant=GPQA Diamond · evaluator=Artificial Analysis · index_version=4.3 | — (obs. 11 Sept 2026) | artificialanalysis.aiT2 | History | |
| #594granite-4-0-h-nano-1bIBM | 26.26% | aa_slug=granite-4-0-h-nano-1b · variant=GPQA Diamond · evaluator=Artificial Analysis · index_version=4.3 | — (obs. 11 Sept 2026) | artificialanalysis.aiT2 | History | |
| #595granite-4-0-350mIBM | 26.06% | aa_slug=granite-4-0-350m · variant=GPQA Diamond · evaluator=Artificial Analysis · index_version=4.3 | — (obs. 11 Sept 2026) | artificialanalysis.aiT2 | History | |
| #596Llama 3.1 8BMeta AI | 25.86% | aa_slug=llama-3-1-instruct-8b · variant=GPQA Diamond · evaluator=Artificial Analysis · index_version=4.3 | — (obs. 11 Sept 2026) | artificialanalysis.aiT2 | History | |
| #597granite-4-0-h-350mIBM | 25.66% | aa_slug=granite-4-0-h-350m · variant=GPQA Diamond · evaluator=Artificial Analysis · index_version=4.3 | — (obs. 11 Sept 2026) | artificialanalysis.aiT2 | History | |
| #598apertus-8b-instructSwiss AI Initiative | 25.56% | aa_slug=apertus-8b-instruct · variant=GPQA Diamond · evaluator=Artificial Analysis · index_version=4.3 | — (obs. 11 Sept 2026) | artificialanalysis.aiT2 | History | |
| #599Llama-3.2-3BMeta AI | 25.45% | aa_slug=llama-3-2-instruct-3b · variant=GPQA Diamond · evaluator=Artificial Analysis · index_version=4.3 | — (obs. 11 Sept 2026) | artificialanalysis.aiT2 | History | |
| #600molmo-7b-dAllen Institute for AI | 24.04% | aa_slug=molmo-7b-d · variant=GPQA Diamond · evaluator=Artificial Analysis · index_version=4.3 | — (obs. 11 Sept 2026) | artificialanalysis.aiT2 | History |
Scores are reported as published, with their evaluation configuration (harness, prompting, judge). The bar is relative to the best score on this page. Results with different configs are not directly comparable — see methodology.
Definition
- Category
- reasoning
Source:AI Atlas curated registry (YAML, versioned in git, every entry carries its source URL)T2observed 9 h agomedium
- Task
- graduate-level science questions
Source:AI Atlas curated registry (YAML, versioned in git, every entry carries its source URL)T2observed 9 h agomedium
- Metric
- accuracy · %
Source:AI Atlas curated registry (YAML, versioned in git, every entry carries its source URL)T2observed 9 h agomedium
- Direction
- Higher is better
- Paper
- https://arxiv.org/abs/2311.12022
Source:AI Atlas curated registry (YAML, versioned in git, every entry carries its source URL)T2observed 9 h agomedium
- Website
Source:AI Atlas curated registry (YAML, versioned in git, every entry carries its source URL)T2observed 9 h agomedium
Each value shows its source, tier and observation time. Missing rows mean no source stated them. How results are recorded →
- Evaluated models
- llama-2-chat-7b, grok-3-mini-reasoning, claude-opus-5-medium, gemma-4-12B, deepseek-v4-flash, grok-4-3-low, gemini-2.0-flash-exp, GPT-5.1-Codex Mini +606(614 total)
As of
Rewind the record: see this entity's attributes exactly as AI Atlas knew them on a given day.
Claim history
6 claims · 6 properties
Categorycategory1
| Value | Valid from → to | Status | Source | Confidence | Extractor |
|---|---|---|---|---|---|
| reasoning | → current | current | AI Atlas curated registry (YAML, versioned in git, every entry carries its source URL)T2 | medium | curated |
Metricmetric1
| Value | Valid from → to | Status | Source | Confidence | Extractor |
|---|---|---|---|---|---|
| accuracy | → current | current | AI Atlas curated registry (YAML, versioned in git, every entry carries its source URL)T2 | medium | curated |
Paperpaper1
| Value | Valid from → to | Status | Source | Confidence | Extractor |
|---|---|---|---|---|---|
| https://arxiv.org/abs/2311.12022 | → current | current | AI Atlas curated registry (YAML, versioned in git, every entry carries its source URL)T2 | medium | curated |
Tasktask1
| Value | Valid from → to | Status | Source | Confidence | Extractor |
|---|---|---|---|---|---|
| graduate-level science questions | → current | current | AI Atlas curated registry (YAML, versioned in git, every entry carries its source URL)T2 | medium | curated |
Unitunit1
| Value | Valid from → to | Status | Source | Confidence | Extractor |
|---|---|---|---|---|---|
| % | → current | current | AI Atlas curated registry (YAML, versioned in git, every entry carries its source URL)T2 | medium | curated |
Websitewebsite1
| Value | Valid from → to | Status | Source | Confidence | Extractor |
|---|---|---|---|---|---|
| https://github.com/idavidrein/gpqa | → current | current | AI Atlas curated registry (YAML, versioned in git, every entry carries its source URL)T2 | medium | curated |
Claims are temporal and append-only: a new observation closes the previous claim (valid_to) instead of overwriting it. Conflicting claims from different sources are kept side by side and flagged — never averaged. Methodology →
| Source | Document | Type | Tier | Last observed | Snapshots |
|---|---|---|---|---|---|
| Artificial Analysis | artificialanalysis.ai/leaderboards/models | leaderboard | T2· Quality secondary | 7 h ago | 1 |
Tier 1 = official/primary, 2 = quality secondary, 3 = community, 4 = unverified. Every snapshot is archived; see all sources and the methodology.