LMCA
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 203 Ergebnissen.
| # | Modell | Organisation | Ergebnis | Nachweis |
|---|---|---|---|---|
| 1 | Claude Fable 5.2 (xhigh) API access | Anthropic | 70,6 % | – |
| 2 | Claude Fable 5.2 (unknown) API access | Anthropic | 70,1 % | – |
| 3 | Claude Fable 5.2 (medium) API access | Anthropic | 70 % | – |
| 4 | Claude Fable 5.2 (high) API access | Anthropic | 69,9 % | – |
| 5 | Claude Fable 5.2 (low) API access | Anthropic | 69,6 % | – |
| 6 | Claude Sonnet 5.5 (max) API access | Anthropic | 68,3 % | – |
| 7 | Claude Opus 5.5 (max) API access | Anthropic | 68,2 % | – |
| 8 | Claude Sonnet 5.5 (xhigh) API access | Anthropic | 66,3 % | – |
| 9 | Claude Opus 5.5 (xhigh) API access | Anthropic | 66,2 % | – |
| 10 | Claude Sonnet 5.5 (high) API access | Anthropic | 65,9 % | – |
| 11 | Claude Opus 5.5 (high) API access | Anthropic | 65,9 % | – |
| 12 | Claude Opus 5.5 (medium) API access | Anthropic | 65,7 % | – |
| 13 | Claude Fable 5.1 (unknown) API access | Anthropic | 65,5 % | – |
| 14 | Claude Sonnet 5.5 (medium) API access | Anthropic | 65,5 % | – |
| 15 | Claude Fable 5.1 (medium) API access | Anthropic | 65,3 % | – |
| 16 | Claude Fable 5.1 (high) API access | Anthropic | 65,1 % | – |
| 17 | Claude Fable 5.1 (xhigh) API access | Anthropic | 65,1 % | – |
| 18 | Claude Opus 5.5 (low) API access | Anthropic | 64,9 % | – |
| 19 | Claude Sonnet 5.5 (low) API access | Anthropic | 64,5 % | – |
| 20 | Claude Opus 5 (xhigh) API access | Anthropic | 64,5 % | – |
| 21 | GPT-6 Astra (xhigh) API access | OpenAI | 64,4 % | – |
| 22 | GPT-6 Astra (unknown thinking) API access | OpenAI | 64,1 % | – |
| 23 | GPT-6 Astra (high) API access | OpenAI | 63,8 % | – |
| 24 | Gemini 4 Argon (high) API access | Google DeepMind | 63,7 % | – |
| 25 | Claude Opus 5 (high) API access | Anthropic | 63,7 % | – |
| 26 | Claude Fable 5.1 (low) API access | Anthropic | 63,5 % | – |
| 27 | Claude Opus 5 (max) API access | Anthropic | 63,3 % | – |
| 28 | Claude Opus 5 (medium) API access | Anthropic | 62,9 % | – |
| 29 | GPT-6 Astra (medium) API access | OpenAI | 62,9 % | – |
| 30 | Gemini 4 Argon (low) API access | Google DeepMind | 62,9 % | – |
| 31 | Gemini 4 Argon (medium) API access | Google DeepMind | 62,9 % | – |
| 32 | GPT-6.1 Sol (max) API access | OpenAI | 62,8 % | – |
| 33 | Claude Opus 5 (low) API access | Anthropic | 62,6 % | – |
| 34 | GPT-6 Astra (low) API access | OpenAI | 62,2 % | – |
| 35 | Claude Fable 5 (high) API access | Anthropic | 61,2 % | – |
| 36 | Claude Fable 5 (xhigh) API access | Anthropic | 61 % | – |
| 37 | Claude Fable 5 (max) API access | Anthropic | 60,3 % | – |
| 38 | Claude Fable 5 (medium) API access | Anthropic | 60,3 % | – |
| 39 | Claude Fable 5 (low) API access | Anthropic | 59,6 % | – |
| 40 | GPT-5.6 Sol (pro, max) API access | OpenAI | 59,2 % | – |
| 41 | GPT-6 Sol (max) API access | OpenAI | 59,1 % | – |
| 42 | GPT-5.6 Sol (xhigh) API access | OpenAI | 58,5 % | – |
| 43 | GPT-5.6 Sol (max) API access | OpenAI | 58,4 % | – |
| 44 | Kimi K3.1 Pro (max) Open weights (restricted use) | Moonshot | 57,6 % | – |
| 45 | Claude Opus 4.8 API access | Anthropic | 57,5 % | – |
| 46 | GPT-5.6 Sol (high) API access | OpenAI | 56,9 % | – |
| 47 | Muse Wonder (unknown) API access | Meta AI | 56,8 % | – |
| 48 | GPT-5.6 Sol (medium) API access | OpenAI | 56,4 % | – |
| 49 | GPT-5.6 Sol (low) API access | OpenAI | 55,9 % | – |
| 50 | Claude Opus 4.6 (max) API access | Anthropic | 55,8 % | – |
| 51 | GLM-5.3 (unknown thinking) Open weights (restricted use) | Z.ai (Zhipu AI) | 55,5 % | – |
| 52 | GPT-5.6 Terra (max) API access | OpenAI | 55 % | – |
| 53 | GPT-5.5 (xhigh) API access | OpenAI | 54,3 % | – |
| 54 | GPT-5.5 Pro (xhigh) API access | OpenAI | 53,9 % | – |
| 55 | Muse Spark 1.3 (unknown) API access | Meta AI | 53,9 % | – |
| 56 | Gemini 3.1 Pro Preview API access | Google DeepMind | 53,8 % | – |
| 57 | Gemini 3.8 Flash (medium) API access | Google DeepMind | 52,9 % | – |
| 58 | Gemini 3.1 Pro Preview API access | Google DeepMind | 52,8 % | – |
| 59 | GPT-5.6 Sol (none) API access | OpenAI | 52,8 % | – |
| 60 | Kimi K3 (max) Open weights (non-commercial) | Moonshot | 52,7 % | – |
| 61 | GPT-5.6 Terra (xhigh) API access | OpenAI | 52,2 % | – |
| 62 | Claude Opus 4.7 (max) API access | Anthropic | 52,2 % | – |
| 63 | GPT-5.4 (xhigh) API access | OpenAI | 52 % | – |
| 64 | Gemini 3.8 Flash (unknown) API access | Google DeepMind | 51,6 % | – |
| 65 | GPT-5.6 Terra (high) API access | OpenAI | 51,4 % | – |
| 66 | Gemini 3.1 Pro Preview API access | Google DeepMind | 51,2 % | – |
| 67 | Gemini 3.8 Flash (low) API access | Google DeepMind | 51,2 % | – |
| 68 | Claude Haiku 5.5 (max) API access | Anthropic | 51,1 % | – |
| 69 | Gemini 3.7 Flash (high) API access | Google DeepMind | 50,4 % | – |
| 70 | Claude Sonnet 5 (high) API access | Anthropic | 50 % | – |
| 71 | Muse Spark 1.1 (high) API access | Meta AI | 49,9 % | – |
| 72 | Claude Sonnet 5 (medium) API access | Anthropic | 49,7 % | – |
| 73 | Claude Sonnet 5 (low) API access | Anthropic | 49,7 % | – |
| 74 | GPT-5.6 Terra (medium) API access | OpenAI | 49,6 % | – |
| 75 | Grok 4.7 (xhigh) API access | xAI | 49,4 % | – |
| 76 | Claude Sonnet 5 (max) API access | Anthropic | 49,3 % | – |
| 77 | GPT-5.6 Terra (low) API access | OpenAI | 49,3 % | – |
| 78 | Gemini 3.7 Flash (medium) API access | Google DeepMind | 49,2 % | – |
| 79 | Claude Sonnet 5 (xhigh) API access | Anthropic | 48,8 % | – |
| 80 | GPT-5.6 Luna (max) API access | OpenAI | 48,5 % | – |
| 81 | GPT-5.6 Luna (xhigh) API access | OpenAI | 48,5 % | – |
| 82 | Grok 4.6 (xhigh) API access | xAI | 48,5 % | – |
| 83 | Muse Spark 1.2 (xhigh) API access | Meta AI | 48,4 % | – |
| 84 | Claude Haiku 5.5 (xhigh) API access | Anthropic | 48 % | – |
| 85 | GPT-5.6 Luna (high) API access | OpenAI | 47,4 % | – |
| 86 | Gemini 3.5 Flash API access | Google DeepMind | 47,2 % | – |
| 87 | Gemini 3.7 Flash (low) API access | Google DeepMind | 47,1 % | – |
| 88 | DeepSeek V4.1 Flash (unknown) Open weights (unrestricted) | DeepSeek | 47 % | – |
| 89 | Claude Sonnet 4.6 (max) API access | Anthropic | 46,5 % | – |
| 90 | GPT-5.6 Terra (none) API access | OpenAI | 46,5 % | – |
| 91 | Qwen3.8 Max (xhigh) API access | Alibaba | 46,2 % | – |
| 92 | GLM-5.2 Open weights (unrestricted) | Z.ai (Zhipu AI) | 45,8 % | – |
| 93 | Claude Haiku 5.5 (medium) API access | Anthropic | 45,6 % | – |
| 94 | DeepSeek V4 Pro 0813 (unknown) Open weights (unrestricted) | DeepSeek | 45,5 % | – |
| 95 | Claude Haiku 5.5 (high) API access | Anthropic | 45,2 % | – |
| 96 | Grok 4.5 (high) API access | xAI | 45,2 % | – |
| 97 | Gemini 3.6 Flash API access | Google DeepMind | 44,9 % | – |
| 98 | Claude Haiku 5.5 (low) API access | Anthropic | 44,8 % | – |
| 99 | GPT-6 Luna (max) API access | OpenAI | 44,5 % | – |
| 100 | Claude Opus 4.5 (high) API access | Anthropic | 44,5 % | – |
