# Mystery Game Puzzles leaderboard

external benchmark

> Every figure below is reproduced as its upstream source published it: nothing is modelled, estimated, interpolated or converted. `Unavailable` means no source published the value — it is never a zero. Each value carries its evidence state and the date it was observed.

Page: https://modelscale.dev/leaderboards/benchmarks/external-mystery-game-puzzles

Mystery Game Puzzles (Epoch AI). Every model the catalog carries a published Mystery Game Puzzles value for, ranked by that value.

## Benchmark record

| Field | Value |
| --- | --- |
| Category | external |
| Measure | Exact-match move answer from text input, minimal agent scaffold |
| Tasks | 100 single-best-move positions from an undisclosed game |
| Difficulty | Unfamiliar game positions, single best move |
| Published by | [Mystery Game Puzzles (Epoch AI)](https://epoch.ai/benchmarks/mystery-game-puzzles) |

Epoch AI's Mystery Game Puzzles benchmark asks models for the single best next move in 100 positions from a puzzle variant of a well-known game whose identity Epoch keeps secret.

## Mystery Game Puzzles ranking

80 models with a published Mystery Game Puzzles value, ordered by that value, highest first. Models the source has not scored on this benchmark are not listed — they are unmeasured, not last.

| Rank | Model | Provider | Exact-match move answer from text input, minimal agent scaffold | Evidence |
| --- | --- | --- | --- | --- |
| 1 | [GPT-6.2 Astra](https://modelscale.dev/models/gpt-6-2-astra) | OpenAI | 95.0 | Observed 2026-11-21 · source benchlm:models |
| 2 | [Claude Fable 5.2](https://modelscale.dev/models/claude-fable-5-2) | Anthropic | 90.0 | Observed 2026-11-21 · source benchlm:models |
| 3 | [GPT-6 Astra](https://modelscale.dev/models/gpt-6-astra) | OpenAI | 84.0 | Observed 2026-11-21 · source benchlm:models |
| 4 | [GPT-6.1 Sol](https://modelscale.dev/models/gpt-6-1-sol) | OpenAI | 80.0 | Observed 2026-11-21 · source benchlm:models |
| 5 | [GPT-6.1 Terra](https://modelscale.dev/models/gpt-6-1-terra) | OpenAI | 74.0 | Observed 2026-11-21 · source benchlm:models |
| 6 | [Gemini 4 Carbon](https://modelscale.dev/models/gemini-4-carbon) | Google | 73.0 | Observed 2026-11-21 · source benchlm:models |
| 7 | [Claude Opus 5.5](https://modelscale.dev/models/claude-opus-5-5) | Anthropic | 71.0 | Observed 2026-11-21 · source benchlm:models |
| 8 | [Claude Sonnet 5.5](https://modelscale.dev/models/claude-sonnet-5-5) | Anthropic | 65.0 | Observed 2026-11-21 · source benchlm:models |
| 8 | [Gemini 4 Hydrogen](https://modelscale.dev/models/gemini-4-hydrogen) | Google | 65.0 | Observed 2026-11-21 · source benchlm:models |
| 10 | [Gemini 4 Argon](https://modelscale.dev/models/gemini-4-argon) | Google | 64.0 | Observed 2026-11-21 · source benchlm:models |
| 11 | [Claude Opus 5](https://modelscale.dev/models/claude-opus-5) | Anthropic | 59.0 | Observed 2026-11-21 · source benchlm:models |
| 12 | [Claude Fable 5.1](https://modelscale.dev/models/claude-fable-5-1) | Anthropic | 58.0 | Observed 2026-11-21 · source benchlm:models |
| 12 | [GPT-5.6 Sol](https://modelscale.dev/models/gpt-5-6-sol) | OpenAI | 58.0 | Observed 2026-11-21 · source benchlm:models |
| 14 | [GPT-5.5](https://modelscale.dev/models/gpt-5-5) | OpenAI | 56.0 | Observed 2026-11-21 · source benchlm:models |
| 14 | [GPT-6 Sol](https://modelscale.dev/models/gpt-6-sol) | OpenAI | 56.0 | Observed 2026-11-21 · source benchlm:models |
| 16 | [Claude Fable 5](https://modelscale.dev/models/claude-fable-5) | Anthropic | 52.0 | Observed 2026-11-21 · source benchlm:models |
| 17 | [Muse Wonder](https://modelscale.dev/models/muse-wonder) | Meta | 51.0 | Observed 2026-11-21 · source benchlm:models |
| 18 | [DeepSeek V4.1 Flash 1017](https://modelscale.dev/models/deepseek-v4-1-flash) | DeepSeek | 49.0 | Observed 2026-11-21 · source benchlm:models |
| 19 | [Gemini 3.8 Flash](https://modelscale.dev/models/gemini-3-8-flash) | Google | 47.0 | Observed 2026-11-21 · source benchlm:models |
| 20 | [DeepSeek V4 Pro 0813](https://modelscale.dev/models/deepseek-v4-pro-max) | DeepSeek | 43.0 | Last good 2026-11-18 · source benchlm:models |
| 21 | [Kimi K3.1 Pro](https://modelscale.dev/models/kimi-3-1-pro) | Moonshot AI | 42.0 | Observed 2026-11-21 · source benchlm:models |
| 22 | [Qwen3.8 Max](https://modelscale.dev/models/qwen3-8-max) | Alibaba | 38.0 | Observed 2026-11-21 · source benchlm:models |
| 23 | [Gemini 3.7 Flash](https://modelscale.dev/models/gemini-3-7-flash) | Google | 37.0 | Observed 2026-11-21 · source benchlm:models |
| 23 | [GPT-5.4](https://modelscale.dev/models/gpt-5-4) | OpenAI | 37.0 | Observed 2026-11-21 · source benchlm:models |
| 25 | [Claude Opus 4.8](https://modelscale.dev/models/claude-opus-4-8) | Anthropic | 36.0 | Observed 2026-11-21 · source benchlm:models |
| 26 | [Claude Sonnet 5](https://modelscale.dev/models/claude-sonnet-5) | Anthropic | 35.0 | Observed 2026-11-21 · source benchlm:models |
| 26 | [GPT-5.6 Terra](https://modelscale.dev/models/gpt-5-6-terra) | OpenAI | 35.0 | Observed 2026-11-21 · source benchlm:models |
| 28 | [DeepSeek V4 Flash 0731](https://modelscale.dev/models/deepseek-v4-flash-max) | DeepSeek | 34.0 | Observed 2026-11-21 · source benchlm:models |
| 28 | [Gemini 3.1 Pro](https://modelscale.dev/models/gemini-3-1-pro) | Google | 34.0 | Observed 2026-11-21 · source benchlm:models |
| 28 | [Grok 4.6](https://modelscale.dev/models/grok-4-6) | xAI | 34.0 | Observed 2026-11-21 · source benchlm:models |
| 31 | [GLM-5.3](https://modelscale.dev/models/glm-5-3) | Z.AI | 33.0 | Observed 2026-11-21 · source benchlm:models |
| 32 | [Gemini 3.5 Flash](https://modelscale.dev/models/gemini-3-5-flash) | Google | 32.0 | Observed 2026-11-21 · source benchlm:models |
| 32 | [Qwen3.7 Max](https://modelscale.dev/models/qwen3-7-max) | Alibaba | 32.0 | Observed 2026-11-21 · source benchlm:models |
| 34 | [Claude Haiku 5.5](https://modelscale.dev/models/claude-haiku-5-5) | Anthropic | 30.0 | Observed 2026-11-21 · source benchlm:models |
| 34 | [Gemini 3.6 Flash](https://modelscale.dev/models/gemini-3-6-flash) | Google | 30.0 | Observed 2026-11-21 · source benchlm:models |
| 36 | [Grok 4.7](https://modelscale.dev/models/grok-4-7) | xAI | 29.0 | Observed 2026-11-21 · source benchlm:models |
| 36 | [o3](https://modelscale.dev/models/o3) | OpenAI | 29.0 | Observed 2026-11-21 · source benchlm:models |
| 38 | [Claude Opus 4.7](https://modelscale.dev/models/claude-opus-4-7) | Anthropic | 28.0 | Observed 2026-11-21 · source benchlm:models |
| 39 | [Kimi K3](https://modelscale.dev/models/kimi-3) | Moonshot AI | 26.0 | Observed 2026-11-21 · source benchlm:models |
| 40 | [Claude Opus 4.6](https://modelscale.dev/models/claude-opus-4-6) | Anthropic | 25.0 | Observed 2026-11-21 · source benchlm:models |
| 40 | [Muse Spark 1.3](https://modelscale.dev/models/muse-spark-1-3) | Meta | 25.0 | Observed 2026-11-21 · source benchlm:models |
| 42 | [GPT-5](https://modelscale.dev/models/gpt-5) | OpenAI | 23.0 | Observed 2026-11-21 · source benchlm:models |
| 42 | [GPT-5 (high)](https://modelscale.dev/models/gpt-5-high) | OpenAI | 23.0 | Last good 2026-11-18 · source benchlm:models |
| 42 | [GPT-5.2](https://modelscale.dev/models/gpt-5-2) | OpenAI | 23.0 | Observed 2026-11-21 · source benchlm:models |
| 45 | [Claude Opus 4.5](https://modelscale.dev/models/claude-opus-4-5) | Anthropic | 22.0 | Observed 2026-11-21 · source benchlm:models |
| 45 | [Qwen3.6-35B-A3B](https://modelscale.dev/models/qwen3-6-35b-a3b) | Alibaba | 22.0 | Observed 2026-11-21 · source benchlm:models |
| 47 | [GPT-5.6 Luna](https://modelscale.dev/models/gpt-5-6-luna) | OpenAI | 21.0 | Observed 2026-11-21 · source benchlm:models |
| 48 | [Gemini 3 Flash](https://modelscale.dev/models/gemini-3-flash) | Google | 20.0 | Observed 2026-11-21 · source benchlm:models |
| 48 | [Nemotron 3 Ultra](https://modelscale.dev/models/nemotron-3-ultra-500b) | NVIDIA | 20.0 | Observed 2026-11-21 · source benchlm:models |
| 48 | [Qwen3.5 Flash](https://modelscale.dev/models/qwen3-5-flash) | Alibaba | 20.0 | Observed 2026-11-21 · source benchlm:models |
| 51 | [Gemini 3.5 Flash-Lite](https://modelscale.dev/models/gemini-3-5-flash-lite) | Google | 19.0 | Observed 2026-11-21 · source benchlm:models |
| 51 | [Qwen 3.6 Max (preview)](https://modelscale.dev/models/qwen3-6-max-preview) | Alibaba | 19.0 | Observed 2026-11-21 · source benchlm:models |
| 53 | [Kimi K2.6](https://modelscale.dev/models/kimi-2-6) | Moonshot AI | 18.0 | Observed 2026-11-21 · source benchlm:models |
| 53 | [Qwen3.5 397B](https://modelscale.dev/models/qwen3-5-397b) | Alibaba | 18.0 | Observed 2026-11-21 · source benchlm:models |
| 55 | [Claude Sonnet 4.5](https://modelscale.dev/models/claude-sonnet-4-5) | Anthropic | 17.0 | Observed 2026-11-21 · source benchlm:models |
| 55 | [Mistral Large 4](https://modelscale.dev/models/mistral-large-4) | Mistral | 17.0 | Observed 2026-11-21 · source benchlm:models |
| 55 | [Qwen3.5 Plus](https://modelscale.dev/models/qwen3-5-plus) | Alibaba | 17.0 | Observed 2026-11-21 · source benchlm:models |
| 55 | [Qwen3.5-122B-A10B](https://modelscale.dev/models/qwen3-5-122b-a10b) | Alibaba | 17.0 | Observed 2026-11-21 · source benchlm:models |
| 55 | [Qwen3.7 Plus](https://modelscale.dev/models/qwen3-7-plus) | Alibaba | 17.0 | Observed 2026-11-21 · source benchlm:models |
| 60 | [Claude Sonnet 4.6](https://modelscale.dev/models/claude-sonnet-4-6) | Anthropic | 16.0 | Observed 2026-11-21 · source benchlm:models |
| 60 | [GPT-5.1](https://modelscale.dev/models/gpt-5-1) | OpenAI | 16.0 | Observed 2026-11-21 · source benchlm:models |
| 62 | [GLM-5.2](https://modelscale.dev/models/glm-5-2) | Z.AI | 15.0 | Observed 2026-11-21 · source benchlm:models |
| 62 | [GPT-5 (medium)](https://modelscale.dev/models/gpt-5-medium) | OpenAI | 15.0 | Last good 2026-11-18 · source benchlm:models |
| 62 | [Qwen3.7 Flash](https://modelscale.dev/models/qwen3-7-flash) | Alibaba | 15.0 | Observed 2026-11-21 · source benchlm:models |
| 65 | [GPT-4o mini](https://modelscale.dev/models/gpt-4o-mini) | OpenAI | 12.0 | Observed 2026-11-21 · source benchlm:models |
| 65 | [Qwen3.6 Plus](https://modelscale.dev/models/qwen3-6-plus) | Alibaba | 12.0 | Observed 2026-11-21 · source benchlm:models |
| 67 | [Qwen3 235B 2507 (Reasoning)](https://modelscale.dev/models/qwen3-235b-2507-reasoning) | Alibaba | 9.0 | Last good 2026-11-18 · source benchlm:models |
| 68 | [GLM-5.3-Flash](https://modelscale.dev/models/glm-5-3-flash) | Z.AI | 8.0 | Observed 2026-11-21 · source benchlm:models |
| 68 | [MiniMax M3](https://modelscale.dev/models/minimax-m3) | MiniMax | 8.0 | Observed 2026-11-21 · source benchlm:models |
| 70 | [GPT-4.1 mini](https://modelscale.dev/models/gpt-4-1-mini) | OpenAI | 7.0 | Observed 2026-11-21 · source benchlm:models |
| 70 | [GPT-5.4 mini](https://modelscale.dev/models/gpt-5-4-mini) | OpenAI | 7.0 | Observed 2026-11-21 · source benchlm:models |
| 70 | [GPT-6 Luna](https://modelscale.dev/models/gpt-6-luna) | OpenAI | 7.0 | Observed 2026-11-21 · source benchlm:models |
| 70 | [o3-mini](https://modelscale.dev/models/o3-mini) | OpenAI | 7.0 | Observed 2026-11-21 · source benchlm:models |
| 70 | [Qwen3.6-27B](https://modelscale.dev/models/qwen3-6-27b) | Alibaba | 7.0 | Observed 2026-11-21 · source benchlm:models |
| 75 | [Inkling-Small](https://modelscale.dev/models/inkling-small) | Thinking Machines Lab | 6.0 | Observed 2026-11-21 · source benchlm:models |
| 76 | [GPT-5.4 nano](https://modelscale.dev/models/gpt-5-4-nano) | OpenAI | 5.0 | Observed 2026-11-21 · source benchlm:models |
| 76 | [o4-mini](https://modelscale.dev/models/o4-mini) | OpenAI | 5.0 | Observed 2026-11-21 · source benchlm:models |
| 76 | [o4-mini (high)](https://modelscale.dev/models/o4-mini-high) | OpenAI | 5.0 | Last good 2026-11-18 · source benchlm:models |
| 76 | [Qwen3 Max](https://modelscale.dev/models/qwen3-max) | Alibaba | 5.0 | Observed 2026-11-21 · source benchlm:models |
| 80 | [GPT-OSS 120B](https://modelscale.dev/models/gpt-oss-120b) | OpenAI | 0.0 | Observed 2026-11-21 · source benchlm:models |

Rows are ordered by the value the source published, highest first. The source does not state whether a higher value is the better result, so this document does not either: for a benchmark that measures a rate of failure, read the table from the bottom.
