Balrog
Each model’s result as Epoch AI’s Benchmarking Hub publishes it, run by Epoch or collected from the source it cites. Facthory has not run them.
Data: Epoch AI, ‘Capabilities & benchmarking’, epoch.ai/benchmarks, licensed CC BY 4.0. Read on 16 Nov 2026.
The chart shows the 25 best of 40 results; the table lists them all.
| # | Model | Organization | Result | Evidence |
|---|---|---|---|---|
| 1 | GPT-6 Astra (max) API access | OpenAI | 68.3% | Log or source |
| 2 | Claude Opus 5 (max) API access | Anthropic | 63.4% | Log or source |
| 3 | GPT-5.6 Sol (max) API access | OpenAI | 60% | Log or source |
| 4 | Gemini 3 Pro Preview API access | Google DeepMind | 58.1% | Log or source |
| 5 | Gemini 3.1 Pro Preview API access | Google DeepMind | 57% | Log or source |
| 6 | GPT-5.6 Terra (max) API access | OpenAI | 53.2% | Log or source |
| 7 | Gemini 3 Flash Preview API access | Google DeepMind | 48.1% | Log or source |
| 8 | GPT-5.6 Luna (max) API access | OpenAI | 45.6% | Log or source |
| 9 | grok-4-0709 API access | xAI | 43.6% | Log or source |
| 10 | Claude Opus 4.5 (no thinking) API access | Anthropic | 43.5% | Log or source |
| 11 | Gemini 2.5 Pro Exp (Mar 2025) API access | Google DeepMind | 43.3% | Log or source |
| 12 | Claude Opus 4.5 (64k thinking) API access | Anthropic | 43% | Log or source |
| 13 | Claude Opus 4.5 (unknown thinking) API access | Anthropic | 43% | Log or source |
| 14 | DeepSeek-R1 Open weights (unrestricted) | DeepSeek | 34.9% | Log or source |
| 15 | gemini-2.5-flash API access | Google DeepMind | 33.5% | Log or source |
| 16 | GPT-5 (minimal) API access | OpenAI | 32.8% | Log or source |
| 17 | Claude 3.5 Sonnet (Oct 2024) API access | Anthropic | 32.6% | Log or source |
| 18 | GPT-4o (May 2024) API access | OpenAI | 32.3% | Log or source |
| 19 | claude-haiku-4-5-20251001 API access | Anthropic | 31.2% | Log or source |
| 20 | claude-haiku-4-5-20251001_1K API access | Anthropic | 31.2% | Log or source |
| 21 | grok-3-beta API access | xAI | 29.5% | Log or source |
| 22 | reka-flash-3 Open weights (unrestricted) | Reka AI | 29.2% | Log or source |
| 23 | Llama-3.1-70B-Instruct Open weights (restricted use) | Meta AI | 27.9% | Log or source |
| 24 | Llama-3.2-90B-Vision-Instruct Open weights (restricted use) | Meta AI | 27.3% | Log or source |
| 25 | Llama-3.3-70B-Instruct Open weights (restricted use) | Meta AI | 23% | Log or source |
| 26 | gemini-1.5-pro-002 API access | Google DeepMind | 21% | Log or source |
| 27 | DeepSeek-R1-Distill-Qwen-32B Open weights (unrestricted) | DeepSeek | 19.5% | Log or source |
| 28 | Claude 3.5 Haiku (Oct 2024) API access | Anthropic | 19.3% | Log or source |
| 29 | Mistral-Nemo-Instruct-2407 Open weights (unrestricted) | Mistral AI | 17.6% | Log or source |
| 30 | gpt-4o-mini-2024-07-18 API access | OpenAI | 17.4% | Log or source |
| 31 | Llama-3.2-11B-Vision-Instruct Open weights (restricted use) | Meta AI | 16.8% | Log or source |
| 32 | qwen2.5-72b-instruct Open weights (unrestricted) | Alibaba | 16.2% | Log or source |
| 33 | Llama-3.1-8B-Instruct Open weights (restricted use) | Meta AI | 15.1% | Log or source |
| 34 | gemini-1.5-flash-002 API access | Google DeepMind | 14.6% | Log or source |
| 35 | Qwen2-VL-72B-Instruct | – | 12.8% | Log or source |
| 36 | phi-4 Open weights (unrestricted) | Microsoft Research | 11.6% | Log or source |
| 37 | Llama-3.2-3B-Instruct Open weights (restricted use) | Meta AI | 10.1% | Log or source |
| 38 | qwen2.5-7b-instruct Open weights (unrestricted) | Alibaba | 7.8% | Log or source |
| 39 | Llama-3.2-1B-Instruct Open weights (restricted use) | Meta AI | 6.6% | Log or source |
| 40 | Qwen2-VL-7B-Instruct | – | 3.7% | Log or source |
