Updated daily at 05:00 UTC
Ranking
Best AI models for Knowledge & reasoning
Broad expert knowledge and multi-step reasoning on hard, closed-answer questions across many fields.
Ranked models
Based on 4 benchmarks
| Rank | Model | Company | Score | Coverage | Released |
|---|---|---|---|---|---|
| 1 | Claude Opus 5.5 | Anthropic | 75.8 | 2 of 4 | 22 Sep 2026 |
| 2 | GPT-6 Astra | OpenAI | 74.2 | 3 of 4 | 3 Sep 2026 |
| 3 | GPT-5.5 Pro | OpenAI | 67.9 | 2 of 4 | 23 Apr 2026 |
| 4 | Muse Spark 1.2 | Meta | 63.1 | 2 of 4 | 5 Aug 2026 |
| 5 | GPT-5.4 Pro | OpenAI | 62.7 | 2 of 4 | 5 Mar 2026 |
| 6 | GPT-6 Sol | OpenAI | 61.4 | 2 of 4 | 22 Sep 2026 |
| 7 | GPT-5.6 Sol | OpenAI | 60.6 | 4 of 4 | 9 Jul 2026 |
| 7 | Gemini 3.8 Flash | Google DeepMind | 60.6 | 2 of 4 | 2 Sep 2026 |
| 9 | Claude Fable 5.1 | Anthropic | 59.5 | 3 of 4 | 1 Sep 2026 |
| 10 | Claude Opus 5 | Anthropic | 59.3 | 4 of 4 | 24 Jul 2026 |
| 11 | Claude Fable 5 | Anthropic | 56.9 | 4 of 4 | 9 Jun 2026 |
| 12 | Gemini 3.7 Flash | Google DeepMind | 51.8 | 2 of 4 | 13 Aug 2026 |
| 13 | GPT-5.6 Terra | OpenAI | 51.6 | 3 of 4 | 9 Jul 2026 |
| 14 | Gemini 3.1 Pro | Google DeepMind | 51.4 | 4 of 4 | 19 Feb 2026 |
| 15 | Kimi K3 | Moonshot AI | 50.2 | 3 of 4 | 16 Jul 2026 |
| 16 | Grok 4.5 | SpaceXAI (xAI) | 49.9 | 3 of 4 | 8 Jul 2026 |
| 17 | Gemini 3 Pro Retired | Google DeepMind | 49.5 | 2 of 4 | 18 Nov 2025 |
| 18 | Muse Spark 1.3 | Meta | 49.1 | 2 of 4 | 2 Sep 2026 |
| 19 | Grok 4.6 | SpaceXAI (xAI) | 48.9 | 4 of 4 | 12 Aug 2026 |
| 19 | DeepSeek-V4-Pro | DeepSeek | 48.9 | 2 of 4 | 24 Apr 2026 |
| 21 | Claude Sonnet 5 | Anthropic | 47.7 | 3 of 4 | 30 Jun 2026 |
| 22 | Claude Opus 4 Retired | Anthropic | 47.3 | 2 of 4 | 22 May 2025 |
| 23 | Gemini 3 Flash | Google DeepMind | 47.0 | 3 of 4 | 17 Dec 2025 |
| 24 | Gemini 3.6 Flash | Google DeepMind | 46.4 | 2 of 4 | 21 Jul 2026 |
| 25 | Claude Opus 4.8 | Anthropic | 46.1 | 4 of 4 | 28 May 2026 |
| 26 | GPT-5.5 | OpenAI | 45.8 | 4 of 4 | 23 Apr 2026 |
| 27 | Qwen3.8-Max | Alibaba Group | 45.6 | 2 of 4 | 2 Aug 2026 |
| 28 | Gemini 3.5 Flash | Google DeepMind | 44.9 | 4 of 4 | 19 May 2026 |
| 29 | DeepSeek-V4-Flash | DeepSeek | 44.0 | 3 of 4 | 24 Apr 2026 |
| 30 | GPT-5.6 Luna | OpenAI | 43.4 | 3 of 4 | 9 Jul 2026 |
| 31 | GPT-5.4 | OpenAI | 42.5 | 3 of 4 | 5 Mar 2026 |
| 32 | o3 Deprecated | OpenAI | 41.7 | 3 of 4 | 16 Apr 2025 |
| 33 | GPT-5.1 | OpenAI | 41.3 | 3 of 4 | 12 Nov 2025 |
| 34 | GLM-5.3 | Z.ai (Zhipu AI) | 41.1 | 2 of 4 | 18 Aug 2026 |
| 35 | Qwen3.6-Max-Preview Deprecated | Alibaba Group | 40.5 | 3 of 4 | 14 Apr 2026 |
| 36 | Claude Opus 4.7 | Anthropic | 40.2 | 4 of 4 | 16 Apr 2026 |
| 37 | GPT-5.2 | OpenAI | 39.0 | 4 of 4 | 11 Dec 2025 |
| 38 | Grok 4 Retired | SpaceXAI (xAI) | 38.4 | 2 of 4 | 9 Jul 2025 |
| 39 | Claude Opus 4.6 | Anthropic | 38.0 | 4 of 4 | 5 Feb 2026 |
| 40 | Gemini 2.5 Pro | Google DeepMind | 37.2 | 3 of 4 | 25 Mar 2025 |
| 41 | GPT-5 Deprecated | OpenAI | 35.4 | 4 of 4 | 7 Aug 2025 |
| 42 | Qwen3.7-Max | Alibaba Group | 35.1 | 4 of 4 | 20 May 2026 |
| 43 | GPT-5.4 nano | OpenAI | 34.9 | 2 of 4 | 17 Mar 2026 |
| 44 | Claude Sonnet 4 Retired | Anthropic | 34.4 | 2 of 4 | 22 May 2025 |
| 45 | GPT-5.4 mini | OpenAI | 34.0 | 2 of 4 | 17 Mar 2026 |
| 46 | Inkling | Thinking Machines Lab | 33.7 | 3 of 4 | 15 Jul 2026 |
| 47 | GPT-5 mini Deprecated | OpenAI | 33.3 | 2 of 4 | 7 Aug 2025 |
| 48 | Kimi K2.7 Code | Moonshot AI | 33.2 | 2 of 4 | 12 Jun 2026 |
| 49 | Grok 4.3 | SpaceXAI (xAI) | 33.1 | 2 of 4 | 17 Apr 2026 |
| 50 | Grok 4.20 | SpaceXAI (xAI) | 32.8 | 2 of 4 | Feb 2026 |
| 51 | GLM-5.2 | Z.ai (Zhipu AI) | 32.7 | 4 of 4 | 16 Jun 2026 |
| 52 | Claude Opus 4.5 | Anthropic | 32.1 | 4 of 4 | 24 Nov 2025 |
| 52 | Qwen3.7-Plus | Alibaba Group | 32.1 | 2 of 4 | 1 Jun 2026 |
| 54 | Claude Sonnet 4.6 | Anthropic | 31.6 | 2 of 4 | 17 Feb 2026 |
| 55 | Gemini 3.1 Flash-Lite | Google DeepMind | 31.1 | 2 of 4 | 3 Mar 2026 |
| 55 | Gemini 3.5 Flash-Lite | Google DeepMind | 31.1 | 2 of 4 | 21 Jul 2026 |
| 57 | Gemini 2.5 Flash | Google DeepMind | 30.9 | 2 of 4 | 17 Apr 2025 |
| 58 | GLM-5.1 | Z.ai (Zhipu AI) | 30.4 | 2 of 4 | 7 Apr 2026 |
| 58 | Qwen3.5-Plus | Alibaba Group | 30.4 | 2 of 4 | 15 Feb 2026 |
| 60 | Qwen3.6-27B | Alibaba Group | 29.3 | 2 of 4 | 22 Apr 2026 |
| 61 | Qwen3.6-35B-A3B | Alibaba Group | 28.5 | 2 of 4 | 16 Apr 2026 |
| 62 | DeepSeek-V3.1 | DeepSeek | 28.3 | 2 of 4 | 21 Aug 2025 |
| 63 | MiniMax-M3 | MiniMax | 28.0 | 3 of 4 | 1 Jun 2026 |
| 64 | o4-mini Deprecated | OpenAI | 26.6 | 3 of 4 | 16 Apr 2025 |
| 65 | Qwen3.5-397B-A17B | Alibaba Group | 26.5 | 2 of 4 | 16 Feb 2026 |
| 66 | Claude Opus 4.1 Retired | Anthropic | 26.4 | 4 of 4 | 5 Aug 2025 |
| 67 | Qwen3.6-Plus | Alibaba Group | 25.8 | 2 of 4 | 2 Apr 2026 |
| 68 | Qwen3.5-Flash | Alibaba Group | 25.6 | 2 of 4 | 23 Feb 2026 |
| 69 | Claude Sonnet 4.5 | Anthropic | 25.1 | 4 of 4 | 29 Sep 2025 |
| 70 | Qwen3.6-Flash | Alibaba Group | 24.8 | 3 of 4 | 16 Apr 2026 |
| 71 | GLM-5 | Z.ai (Zhipu AI) | 24.6 | 2 of 4 | 12 Feb 2026 |
| 72 | Qwen3-235B-A22B | Alibaba Group | 23.3 | 2 of 4 | 29 Apr 2025 |
| 73 | Gemma 4 31B | Google DeepMind | 23.1 | 2 of 4 | 31 Mar 2026 |
| 74 | Kimi K2.6 | Moonshot AI | 22.8 | 3 of 4 | 20 Apr 2026 |
| 75 | DeepSeek-V3.2 | DeepSeek | 21.7 | 2 of 4 | 1 Dec 2025 |
| 76 | o1 Deprecated | OpenAI | 21.6 | 3 of 4 | 5 Dec 2024 |
| 77 | Qwen3-235B-A22B-Thinking-2507 | Alibaba Group | 20.9 | 2 of 4 | 25 Jul 2025 |
| 78 | Qwen3.5-35B-A3B | Alibaba Group | 20.0 | 2 of 4 | 24 Feb 2026 |
| 79 | Claude Haiku 4.5 | Anthropic | 19.8 | 2 of 4 | 15 Oct 2025 |
| 80 | GLM-4.7 | Z.ai (Zhipu AI) | 19.2 | 2 of 4 | 22 Dec 2025 |
| 81 | Qwen3.5-9B | Alibaba Group | 18.1 | 2 of 4 | 2 Mar 2026 |
| 82 | Gemma 4 26B A4B | Google DeepMind | 18.0 | 2 of 4 | 31 Mar 2026 |
| 83 | Qwen3-Max Deprecated | Alibaba Group | 16.6 | 2 of 4 | 5 Sep 2025 |
| 84 | GPT-5 nano Deprecated | OpenAI | 16.3 | 2 of 4 | 7 Aug 2025 |
| 85 | gpt-oss-120b | OpenAI | 16.1 | 3 of 4 | 5 Aug 2025 |
| 86 | Llama 4 Maverick | Meta | 16.0 | 2 of 4 | 5 Apr 2025 |
| 87 | GPT-4.1 | OpenAI | 14.5 | 3 of 4 | 14 Apr 2025 |
| 88 | o3-mini Deprecated | OpenAI | 14.1 | 3 of 4 | 31 Jan 2025 |
| 89 | GPT-4.1 mini | OpenAI | 13.5 | 2 of 4 | 14 Apr 2025 |
| 90 | Qwen3-30B-A3B-Instruct-2507 | Alibaba Group | 13.2 | 2 of 4 | 30 Jul 2025 |
| 91 | Qwen3-30B-A3B-Thinking-2507 | Alibaba Group | 13.1 | 2 of 4 | 31 Jul 2025 |
| 92 | Llama 3.3 70B | Meta | 12.2 | 2 of 4 | 6 Dec 2024 |
| 93 | Qwen3-14B | Alibaba Group | 10.7 | 2 of 4 | 29 Apr 2025 |
| 94 | Qwen3-32B | Alibaba Group | 10.2 | 2 of 4 | 29 Apr 2025 |
| 95 | GPT-4 Deprecated | OpenAI | 10.0 | 2 of 4 | 14 Mar 2023 |
| 96 | GPT-4o | OpenAI | 9.8 | 3 of 4 | 13 May 2024 |
| 97 | Claude 3 Opus Retired | Anthropic | 9.4 | 3 of 4 | 4 Mar 2024 |
| 98 | Qwen3-30B-A3B | Alibaba Group | 9.3 | 2 of 4 | 29 Apr 2025 |
| 99 | gpt-oss-20b | OpenAI | 8.5 | 2 of 4 | 5 Aug 2025 |
| 100 | GPT-4 Turbo Deprecated | OpenAI | 7.6 | 3 of 4 | 6 Nov 2023 |
| 101 | Gemma 3 27B | Google DeepMind | 7.2 | 2 of 4 | 10 Mar 2025 |
| 102 | GPT-3.5 Turbo Deprecated | OpenAI | 5.7 | 2 of 4 | 1 Mar 2023 |
| 103 | Qwen3-8B | Alibaba Group | 5.2 | 2 of 4 | 29 Apr 2025 |
| 104 | GPT-4o mini | OpenAI | 4.1 | 3 of 4 | 18 Jul 2024 |
| 105 | Qwen2.5-7B-Instruct | Alibaba Group | 3.8 | 2 of 4 | 19 Sep 2024 |
| 106 | Command R+ (08-2024) | Cohere | 3.4 | 2 of 4 | 30 Aug 2024 |
| 107 | Llama 3.1 8B | Meta | 3.2 | 2 of 4 | 23 Jul 2024 |
| 108 | Gemma 3 12B | Google DeepMind | 2.6 | 2 of 4 | 10 Mar 2025 |
| 109 | Gemma 3 4B | Google DeepMind | 1.6 | 2 of 4 | 10 Mar 2025 |
Not ranked yet
These models have independent or third-party results on too few of the 4 benchmarks to rank.
