EnigmaEval
Multi-step puzzle-hunt problems that need creative leaps.
| Category | Reasoning |
|---|---|
| Metric | Accuracy |
| Source | Epoch AI (evaluator) |
| License | CC BY 4.0 |
| Benchmark page | scale.com |
| Status | Scored |
| Weight | 0.8 within Reasoning |
| Fit | Discrimination 1.34, noise 0.53 |
Results
| # | Model | Provider | Result | Date |
|---|---|---|---|---|
| 1 | Claude Fable 5 High | Anthropic | 39.3% | 2026-11-11 |
| 2 | GPT-5.6 Sol High | OpenAI | 37.1% | 2026-11-11 |
| 3 | Gemini 3.1 Pro Preview | 36.8% | 2026-11-11 | |
| 4 | Gemini 3.5 Flash High | 25.4% | 2026-11-11 | |
| 5 | GPT-5.4 Pro | OpenAI | 23.8% | 2026-11-11 |
| 6 | Claude Opus 4.8 Extra High | Anthropic | 23.5% | 2026-11-11 |
| 7 | GPT-5 Pro | OpenAI | 18.8% | 2026-11-11 |
| 8 | Gemini 3 Pro Preview | 18.2% | 2026-11-11 | |
| 9 | GPT-5.4 Extra High | OpenAI | 16.0% | 2026-11-11 |
| 10 | o3 Medium | OpenAI | 13.1% | 2026-11-11 |
| 11 | Claude Opus 4.5 | Anthropic | 11.9% | 2026-11-11 |
| 12 | o3 High | OpenAI | 11.9% | 2026-11-11 |
| 13 | GPT-5.1 | OpenAI | 11.2% | 2026-11-11 |
| 14 | GPT-5 | OpenAI | 10.5% | 2026-11-11 |
| 15 | GPT-5.2 | OpenAI | 10.4% | 2026-11-11 |
| 16 | o4-mini High | OpenAI | 9.2% | 2026-11-11 |
| 17 | GPT-5 Mini | OpenAI | 8.2% | 2026-11-11 |
| 18 | Claude Opus 4.6 Max | Anthropic | 7.6% | 2026-11-11 |
| 19 | Claude Opus 4.1 | Anthropic | 7.2% | 2026-11-11 |
| 20 | Claude Opus 4.6 | Anthropic | 6.8% | 2026-11-11 |
| 21 | o4-mini Medium | OpenAI | 6.8% | 2026-11-11 |
| 22 | o1-pro | OpenAI | 6.1% | 2026-11-11 |
| 23 | Claude Sonnet 4.5 | Anthropic | 6.0% | 2026-11-11 |
| 24 | o1 | OpenAI | 5.7% | 2026-11-11 |
| 25 | Claude Opus 4 | Anthropic | 5.6% | 2026-11-11 |
| 26 | Gemini 2.5 Pro Preview 06-05 | 5.6% | 2026-11-11 | |
| 27 | Claude Sonnet 3.7 | Anthropic | 4.2% | 2026-11-11 |
| 28 | Kimi K2.5 | Moonshot AI | 3.4% | 2026-11-11 |
| 29 | GPT-4.5 | OpenAI | 3.2% | 2026-11-11 |
| 30 | Claude Sonnet 4 | Anthropic | 3.1% | 2026-11-11 |
| 31 | Gemini 3.1 Flash Lite | 3.0% | 2026-11-11 | |
| 32 | Gemini 2.5 Flash Preview 05-20 | 2.7% | 2026-11-11 | |
| 33 | GPT-4.1 | OpenAI | 2.2% | 2026-11-11 |
| 34 | GPT-5.1 Non-reasoning | OpenAI | 1.9% | 2026-11-11 |
| 35 | Gemini 2.0 Flash Thinking (Jan 2025) | 1.1% | 2026-11-11 | |
| 36 | Claude Sonnet 3.5 v2 | Anthropic | 0.9% | 2026-11-11 |
| 37 | Pixtral Large (latest) | Mistral | 0.8% | 2026-11-11 |
| 38 | Claude 3 Opus | Anthropic | 0.8% | 2026-11-11 |
| 39 | GPT-4o (2024-11-20) | OpenAI | 0.8% | 2026-11-11 |
| 40 | Gemini 2.0 Pro | 0.7% | 2026-11-11 | |
| 41 | Gemini 2.0 Flash | 0.6% | 2026-11-11 | |
| 42 | Llama 4 Maverick | Meta | 0.6% | 2026-11-11 |
| 43 | Llama-3.2-90B-Vision-Instruct | Meta | 0.4% | 2026-11-11 |