SimpleBench
Das Ergebnis jedes Modells, so, wie der Benchmarking Hub von Epoch AI es veröffentlicht, von Epoch gemessen oder aus der genannten Quelle gesammelt. Facthory hat sie nicht selbst gemessen.
Daten: Epoch AI, „Capabilities & benchmarking“, epoch.ai/benchmarks, lizenziert unter CC BY 4.0. Gelesen am 16.11.2026.
Das Diagramm zeigt die 25 besten, die Tabelle die 100 besten von 124 Ergebnissen.
| # | Modell | Organisation | Ergebnis | Nachweis |
|---|---|---|---|---|
| 1 | Claude Fable 5.2 (unknown) API access | Anthropic | 94,9 % | – |
| 2 | Gemini 4 Carbon (unknown) API access | Google DeepMind | 92,6 % | – |
| 3 | Gemini 4 Argon (unknown) API access | Google DeepMind | 91,7 % | – |
| 4 | Claude Opus 5.5 (unknown) API access | Anthropic | 88,4 % | – |
| 5 | Claude Fable 5.1 (unknown) API access | Anthropic | 86,6 % | – |
| 6 | GPT-6 Astra (pro, max) API access | OpenAI | 86,5 % | – |
| 7 | GPT-6 Astra (unknown thinking) API access | OpenAI | 83,6 % | – |
| 8 | GPT-6.1 Sol (unknown thinking) API access | OpenAI | 82,9 % | – |
| 9 | Gemini 3.8 Flash (unknown) API access | Google DeepMind | 82,4 % | – |
| 10 | Claude Fable 5 (max) API access | Anthropic | 81,9 % | – |
| 11 | Muse Spark 1.3 (unknown) API access | Meta AI | 81,8 % | – |
| 12 | Claude Opus 5 (unknown) API access | Anthropic | 80,6 % | – |
| 13 | Gemini 3.1 Pro Preview API access | Google DeepMind | 79,6 % | – |
| 14 | GPT-5.5 Pro (high) API access | OpenAI | 76,9 % | – |
| 15 | GPT-5.5 Pro (unknown thinking) API access | OpenAI | 76,9 % | – |
| 16 | Gemini 3.5 Flash API access | Google DeepMind | 76,7 % | – |
| 17 | Gemini 3 Pro Preview API access | Google DeepMind | 76,4 % | – |
| 18 | Claude Sonnet 5.5 (unknown) API access | Anthropic | 75,9 % | – |
| 19 | Grok 4.6 (unknown) API access | xAI | 75,9 % | – |
| 20 | Muse Spark 1.2 API access | Meta AI | 74,5 % | – |
| 21 | GPT-5.4 Pro API access | OpenAI | 74,1 % | – |
| 22 | GPT-6 Sol (unknown thinking) API access | OpenAI | 73,1 % | – |
| 23 | Muse Ember (unknown) API access | Meta AI | 72,7 % | – |
| 24 | GPT-6.1 Terra (unknown thinking) API access | OpenAI | 72,2 % | – |
| 25 | GPT-5.6 Sol (pro, xhigh) API access | OpenAI | 71,7 % | – |
| 26 | GPT-5.6 Sol (pro, unknown thinking) API access | OpenAI | 71,7 % | – |
| 27 | Qwen3.7 Max API access | Alibaba | 70,4 % | – |
| 28 | Grok 4.5 (unknown thinking) API access | xAI | 70 % | – |
| 29 | GPT-5.5 (unknown thinking) API access | OpenAI | 69 % | – |
| 30 | Claude Opus 4.6 (no thinking) API access | Anthropic | 67,6 % | – |
| 31 | DeepSeek V4.1 Flash (unknown) Open weights (unrestricted) | DeepSeek | 66,7 % | – |
| 32 | GLM-5.3 (unknown thinking) Open weights (restricted use) | Z.ai (Zhipu AI) | 66,2 % | – |
| 33 | Claude Opus 4.8 API access | Anthropic | 64,8 % | – |
| 34 | GPT-5.6 Sol (unknown thinking) API access | OpenAI | 64,8 % | – |
| 35 | GPT-5.6 Sol (xhigh) API access | OpenAI | 64,8 % | – |
| 36 | Qwen 3.6 Max (Preview) | Alibaba | 63 % | – |
| 37 | Gemini 2.5 Pro Preview (Jun 2025) API access | Google DeepMind | 62,4 % | – |
| 38 | Claude Opus 4.5 (no thinking) API access | Anthropic | 62 % | – |
| 39 | Claude Opus 4.5 (unknown thinking) API access | Anthropic | 62 % | – |
| 40 | Claude Opus 4.7 (unknown) API access | Anthropic | 61,7 % | – |
| 41 | GPT-5 Pro API access | OpenAI | 61,6 % | – |
| 42 | GPT-5 Pro API access | OpenAI | 61,6 % | – |
| 43 | DeepSeek V4 Flash 0731 (unknown) Open weights (unrestricted) | DeepSeek | 61,1 % | – |
| 44 | Gemini 3 Flash Preview API access | Google DeepMind | 61,1 % | – |
| 45 | Kimi K3 (max) Open weights (non-commercial) | Moonshot | 60,7 % | – |
| 46 | Claude Sonnet 5 (unknown thinking) API access | Anthropic | 60,6 % | – |
| 47 | grok-4-0709 API access | xAI | 60,5 % | – |
| 48 | Qwen3.8 27B Open weights (unrestricted) | Alibaba | 60,2 % | – |
| 49 | claude-opus-4-1-20250805 API access | Anthropic | 60 % | – |
| 50 | Claude Opus 4.1 (unknown thinking) API access | Anthropic | 60 % | – |
| 51 | Claude Opus 4 API access | Anthropic | 58,8 % | – |
| 52 | Claude Opus 4 API access | Anthropic | 58,8 % | – |
| 53 | GLM-5.2 Open weights (unrestricted) | Z.ai (Zhipu AI) | 58,8 % | – |
| 54 | Kimi K2.7 Code Open weights (unrestricted) | Moonshot | 57,9 % | – |
| 55 | GPT-5.2 Pro API access | OpenAI | 57,4 % | – |
| 56 | GPT-5 (high) API access | OpenAI | 56,7 % | – |
| 57 | grok-4-1-fast-non-reasoning API access | xAI | 56 % | – |
| 58 | grok-4-1-fast-reasoning API access | xAI | 56 % | – |
| 59 | GLM-5.1 Open weights (unrestricted) | Z.ai (Zhipu AI) | 55,1 % | – |
| 60 | Claude Sonnet 4.5 (12k thinking) API access | Anthropic | 54,3 % | – |
| 61 | Claude Sonnet 4.5 (unknown thinking) API access | Anthropic | 54,3 % | – |
| 62 | GLM-5 Open weights (unrestricted) | Z.ai (Zhipu AI) | 53,2 % | – |
| 63 | GPT-5.1 (high) API access | OpenAI | 53,2 % | – |
| 64 | o3 (high) API access | OpenAI | 53,1 % | – |
| 65 | DeepSeek-V3.2-Speciale Open weights (unrestricted) | DeepSeek | 52,6 % | – |
| 66 | Gemini 2.5 Pro Exp (Mar 2025) API access | Google DeepMind | 51,6 % | – |
| 67 | Gemini 2.5 Pro Preview (Mar 2025) API access | Google DeepMind | 51,6 % | – |
| 68 | Inkling Open weights (unrestricted) | Thinking Machines | 50 % | – |
| 69 | GPT-5.6 Terra (unknown thinking) API access | OpenAI | 48,9 % | – |
| 70 | GPT-5.6 Terra (xhigh) API access | OpenAI | 48,9 % | – |
| 71 | GLM-4.7 (Novita) Open weights (unrestricted) | Z.ai (Zhipu AI) | 47,7 % | – |
| 72 | GLM-4.7 Open weights (unrestricted) | Z.ai (Zhipu AI) | 47,7 % | – |
| 73 | GPT-5.6 Luna (unknown thinking) API access | OpenAI | 46,8 % | – |
| 74 | GPT-5.6 Luna (xhigh) API access | OpenAI | 46,8 % | – |
| 75 | Kimi K2.5 Open weights (unrestricted) | Moonshot | 46,8 % | – |
| 76 | Claude 3.7 Sonnet (12k thinking) API access | Anthropic | 46,4 % | – |
| 77 | Claude 3.7 Sonnet (unknown thinking) API access | Anthropic | 46,4 % | – |
| 78 | DeepSeek v4 (unknown) Open weights (unrestricted) | DeepSeek | 46,3 % | – |
| 79 | GPT-5.2 (high) API access | OpenAI | 45,8 % | – |
| 80 | MiniMax-M3 Open weights (restricted use) | MiniMax | 45,8 % | – |
| 81 | GPT-5.2 (unknown thinking) API access | OpenAI | 45,8 % | – |
| 82 | claude-sonnet-4-20250514_12K API access | Anthropic | 45,5 % | – |
| 83 | Claude Sonnet 4 (unknown thinking) API access | Anthropic | 45,5 % | – |
| 84 | claude-3-7-sonnet-20250219 API access | Anthropic | 44,9 % | – |
| 85 | nemotron-3-ultra Open weights (unrestricted) | Nvidia | 41,7 % | – |
| 86 | o1-preview API access | OpenAI | 41,7 % | – |
| 87 | Claude 3.5 Sonnet (Oct 2024) API access | Anthropic | 41,4 % | – |
| 88 | gemini-2.5-flash API access | Google DeepMind | 41,2 % | – |
| 89 | DeepSeek-R1 (May 2025) Open weights (unrestricted) | DeepSeek | 40,8 % | – |
| 90 | o1 (high) API access | OpenAI | 40,1 % | – |
| 91 | DeepSeek-V3.1 Open weights (unrestricted) | DeepSeek | 40 % | – |
| 92 | Kimi K2 Thinking Open weights (restricted use) | Moonshot | 39,6 % | – |
| 93 | o4-mini (high) API access | OpenAI | 38,7 % | – |
| 94 | o1 (medium) API access | OpenAI | 36,7 % | – |
| 95 | grok-3 API access | xAI | 36,1 % | – |
| 96 | Qwen 3.6 Flash (2026-04-16) | Alibaba | 35,2 % | – |
| 97 | MiniMax-M2.1 Open weights (restricted use) | MiniMax | 34,7 % | – |
| 98 | GPT-4.5 Preview (Feb 2025) API access | OpenAI | 34,5 % | – |
| 99 | gemini-exp-1206 API access | Google DeepMind,Google | 31,1 % | – |
| 100 | qwen3-235b-a22b Open weights (unrestricted) | Alibaba | 31 % | – |
