Mystery game puzzles leaderboard
- GPT-6 Astra leads
- 84.0%
- models measured
- 73
- lead over #2
- +4.0 pts
- median
- 23.0%
Leader on Mystery game puzzles · read October 21
On October 21, 2026, 73 of our catalog's models had a Mystery game puzzles result, and OpenAI GPT-6 Astra led with 84.0%.
Epoch AI collects its results; its benchmarking hub describes the tasks and links the authors.
Role in the index
We collect this benchmark and show it as supporting evidence. It is not part of the Vibe Coding Index, because it measures something other than coding and the index is about coding. Read the methodology
| # | Model | Result | Evidence | Coding index | Output $/Mtok |
|---|---|---|---|---|---|
| 1 | GPT-6 AstraOpenAI | 84.0% | 88.7 | $50.00 | |
| 2 | GPT-6.1 SolOpenAI | 80.0% | 88.7 | $10.00 | |
| 3 | Claude Opus 5.5Anthropic | 71.0% | 92.2 | $20.00 | |
| 4 | Claude Sonnet 5.5Anthropic | 65.0% | 88.9 | $10.00 | |
| 5 | Gemini 4 ArgonGoogle | 64.0% | 80.5 | $10.00 | |
| 6 | Claude Opus 5Anthropic | 59.0% | 81.8 | $25.00 | |
| 7 | Claude Fable 5.1Anthropic | 58.0% | 89.1 | $50.00 | |
| 8 | GPT-5.6 SolOpenAI | 58.0% | 76.7 | $20.00 | |
| 9 | GPT-5.5OpenAI | 56.0% | 67.8 | $30.00 | |
| 10 | GPT-6 SolOpenAI | 56.0% | 82.8 | $10.00 | |
| 11 | Claude Fable 5Anthropic | 52.0% | 82.1 | $50.00 | |
| 12 | Gemini 3.8 FlashGoogle | 47.0% | 70.5 | $4.125 | |
| 13 | DeepSeek V4 Pro 0813DeepSeek | 43.0% | 54.6 | $1.98 | |
| 14 | DeepSeek V4.1 FlashDeepSeek | 43.0% | 77.2 | $0.525 | |
| 15 | Qwen3.8 Max (0902)Qwen | 38.0% | 63.5 | $6.00 | |
| 16 | Gemini 3.7 FlashGoogle | 37.0% | 69.9 | $3.75 | |
| 17 | GPT-5.4OpenAI | 37.0% | 37.7 | $15.00 | |
| 18 | Claude Opus 4.8Anthropic | 36.0% | 66.0 | $25.00 | |
| 19 | Claude Sonnet 5Anthropic | 35.0% | 63.5 | $10.00 | |
| 20 | GPT-5.6 TerraOpenAI | 35.0% | 69.0 | $12.00 | |
| 21 | DeepSeek V4 Flash 0731DeepSeek | 34.0% | Not benchmarked | $1.32 | |
| 22 | Gemini 3.1 Pro PreviewGoogle | 34.0% | 35.3 | $12.00 | |
| 23 | Grok 4.6xAI | 34.0% | 76.4 | $6.00 | |
| 24 | GLM 5.3Z.ai | 33.0% | 73.1 | $1.40 | |
| 25 | Gemini 3.5 FlashGoogle | 32.0% | 42.8 | $9.00 | |
| 26 | Qwen3.7 MaxQwen | 32.0% | 45.1 | $4.425 | |
| 27 | Claude Haiku 5.5Anthropic | 30.0% | 75.9 | $0.50 | |
| 28 | Gemini 3.6 FlashGoogle | 30.0% | 48.3 | $3.75 | |
| 29 | o3OpenAI | 29.0% | 70.7 | $8.00 | |
| 30 | Grok 4.7xAI | 29.0% | 75.0 | $6.00 | |
| 31 | Claude Opus 4.7Anthropic | 28.0% | 51.1 | $25.00 | |
| 32 | Gemini 3 Flash PreviewGoogle | 26.0% | 75.8 | $3.00 | |
| 33 | Kimi K3Moonshot AI | 26.0% | 61.8 | $11.936 | |
| 34 | Claude Opus 4.6Anthropic | 25.0% | 66.9 | $25.00 | |
| 35 | Muse Spark 1.3Meta | 25.0% | 79.7 | $4.25 | |
| 36 | GPT-5OpenAI | 23.0% | 72.2 | $10.00 | |
| 37 | GPT-5.2OpenAI | 23.0% | 58.9 | $14.00 | |
| 38 | Claude Opus 4.5Anthropic | 22.0% | 66.8 | $25.00 | |
| 39 | Qwen3.6 35B A3BQwen | 22.0% | Not benchmarked | $0.95 | |
| 40 | Claude Opus 4.1Anthropic | 21.0% | 27.2 | $75.00 | |
| 41 | GPT-5.6 LunaOpenAI | 21.0% | 61.5 | $1.20 | |
| 42 | Nemotron 3 UltraNVIDIA | 20.0% | Not benchmarked | $2.40 | |
| 43 | Qwen3.5-FlashQwen | 20.0% | 6.2 | $0.26 | |
| 44 | Gemini 3.5 Flash LiteGoogle | 19.0% | 33.6 | $2.50 | |
| 45 | GPT-5.1OpenAI | 19.0% | 53.3 | $10.00 | |
| 46 | GLM 5.2Z.ai | 19.0% | 57.7 | $7.00 | |
| 47 | Kimi K2.6Moonshot AI | 18.0% | 44.2 | $3.41 | |
| 48 | Qwen3.5 397B A17BQwen | 18.0% | 28.6 | $2.34 | |
| 49 | Qwen3.6 FlashQwen | 18.0% | Not benchmarked | $1.125 | |
| 50 | Claude Sonnet 4.5Anthropic | 17.0% | 59.1 | $15.00 | |
| 51 | DeepSeek V4 Pro 0423DeepSeek | 17.0% | 35.3 | $1.9105 | |
| 52 | Mistral Large 4Mistral AI | 17.0% | 55.3 | $4.18 | |
| 53 | Qwen3.5 Plus 2026-04-20Qwen | 17.0% | Not benchmarked | $1.80 | |
| 54 | Qwen3.5-122B-A10BQwen | 17.0% | 22.9 | $2.08 | |
| 55 | Qwen3.7 PlusQwen | 17.0% | Not benchmarked | $1.28 | |
| 56 | Claude Sonnet 4.6Anthropic | 16.0% | 57.8 | $15.00 | |
| 57 | Qwen3.7 FlashQwen | 15.0% | Not benchmarked | $0.13 | |
| 58 | GPT-4o-mini (2024-07-18)OpenAI | 12.0% | 3.6 | $0.60 | |
| 59 | Qwen3.6 PlusQwen | 12.0% | 37.4 | $1.95 | |
| 60 | GPT-5.4 MiniOpenAI | 11.0% | 28.2 | $4.50 | |
| 61 | GPT-5 MiniOpenAI | 10.0% | 59.8 | $2.00 | |
| 62 | GPT-5 NanoOpenAI | 9.0% | 34.8 | $0.40 | |
| 63 | GPT-5.4 NanoOpenAI | 9.0% | Not benchmarked | $1.25 | |
| 64 | Qwen3 235B A22B Thinking 2507Qwen | 9.0% | Not benchmarked | $3.50 | |
| 65 | MiniMax M3MiniMax | 8.0% | 57.8 | $1.20 | |
| 66 | GLM 5.3 FlashZ.ai | 8.0% | 71.0 | $0.50 | |
| 67 | GPT-4.1 MiniOpenAI | 7.0% | 28.5 | $1.60 | |
| 68 | GPT-6 LunaOpenAI | 7.0% | 71.4 | $0.50 | |
| 69 | o3 MiniOpenAI | 7.0% | 52.1 | $4.40 | |
| 70 | Qwen3.6 27BQwen | 7.0% | Not benchmarked | $2.00 | |
| 71 | T Inkling SmallThinking Machines | 6.0% | 29.9 | $1.20 | |
| 72 | o4 MiniOpenAI | 5.0% | 68.6 | $4.40 | |
| 73 | gpt-oss-120bOpenAI | 2.0% | 34.5 | $0.60 |
Only models in our catalog are listed, and each appears once with its best published run. 197 rows were read from the source on October 21, 2026, and rows that match no catalog model are left out.