Extended NYT Connections
940 New York Times Connections word-grouping puzzles with up to four extra decoy words added, testing verbal association and constraint solving. Run by Lech Mazur.
| Category | Reasoning |
|---|---|
| Metric | Puzzle score (%) |
| Source | Lech Mazur (original) |
| License | None stated |
| Benchmark page | github.com |
| Status | Scored |
| Weight | 0.9 within Reasoning |
| Fit | Discrimination 1.99, noise 0.49 |
Results
| # | Model | Provider | Result | Date |
|---|---|---|---|---|
| 1 | Gemini 4 Carbon High | 99.3% | 2026-11-10 | |
| 2 | Gemini 4 Argon High | 98.6% | 2026-11-10 | |
| 3 | GPT-6 Astra Max | OpenAI | 98.4% | 2026-11-10 |
| 4 | GPT-6 Astra Extra High | OpenAI | 98.1% | 2026-11-10 |
| 5 | GPT-6.1 Sol Max | OpenAI | 97.8% | 2026-11-10 |
| 6 | GPT-6 Astra High | OpenAI | 97.7% | 2026-11-10 |
| 7 | Gemini 3.1 Pro Preview | 97.4% | 2026-11-10 | |
| 8 | Gemini 3.8 Flash High | 97.4% | 2026-11-10 | |
| 9 | Claude Fable 5.2 High | Anthropic | 97.3% | 2026-11-10 |
| 10 | GPT-5.5 Extra High | OpenAI | 96.2% | 2026-11-10 |
| 11 | GPT-6.1 Sol High | OpenAI | 95.5% | 2026-11-10 |
| 12 | GPT-5.5 High | OpenAI | 95.2% | 2026-11-10 |
| 13 | Gemini 3 Pro Preview | 94.4% | 2026-11-10 | |
| 14 | Claude Opus 5 Extra High | Anthropic | 94.3% | 2026-11-10 |
| 15 | Gemini 3.7 Flash | 94.0% | 2026-11-10 | |
| 16 | GPT-5.6 Sol Extra High | OpenAI | 93.8% | 2026-11-10 |
| 17 | Kimi K3 | Moonshot AI | 93.6% | 2026-11-10 |
| 18 | Muse Wonder High | Meta | 93.4% | 2026-11-10 |
| 19 | Kimi K3.1 Pro High | Moonshot AI | 93.4% | 2026-11-10 |
| 20 | DeepSeek V4.1 Flash 1017 High | DeepSeek | 92.8% | 2026-11-10 |
| 21 | Claude Fable 5 High | Anthropic | 92.7% | 2026-11-10 |
| 22 | Gemini 3.5 Flash | 92.6% | 2026-11-10 | |
| 23 | GPT-5.5 Medium | OpenAI | 92.4% | 2026-11-10 |
| 24 | Claude Opus 5 High | Anthropic | 92.2% | 2026-11-10 |
| 25 | Claude Opus 4.6 High | Anthropic | 92.1% | 2026-11-10 |
| 26 | GPT-5.4 Extra High | OpenAI | 91.3% | 2026-11-10 |
| 27 | DeepSeek V4 Pro High | DeepSeek | 91.3% | 2026-11-10 |
| 28 | Claude Opus 4.8 Extra High | Anthropic | 91.1% | 2026-11-10 |
| 29 | GPT-5.6 Sol High | OpenAI | 91.0% | 2026-11-10 |
| 30 | GPT-5.4 High | OpenAI | 90.6% | 2026-11-10 |
| 31 | GPT-6 Sol High | OpenAI | 90.1% | 2026-11-10 |
| 32 | Claude Fable 5.1 High | Anthropic | 90.0% | 2026-11-10 |
| 33 | DeepSeek V4 Flash | DeepSeek | 89.6% | 2026-11-10 |
| 34 | DeepSeek V4.1 Flash High | DeepSeek | 89.6% | 2026-11-10 |
| 35 | Gemini 3.6 Flash | 89.0% | 2026-11-10 | |
| 36 | Claude Opus 5.5 High | Anthropic | 88.5% | 2026-11-10 |
| 37 | Qwen3.8 Max | Alibaba | 88.3% | 2026-11-10 |
| 38 | Claude Opus 4.8 High | Anthropic | 88.3% | 2026-11-10 |
| 39 | GPT-5.4 Medium | OpenAI | 87.8% | 2026-11-10 |
| 40 | Claude Opus 5.5 Medium | Anthropic | 87.8% | 2026-11-10 |
| 41 | Grok 4.1 Fast (Reasoning) | xAI | 87.4% | 2026-11-10 |
| 42 | Kimi K2.6 | Moonshot AI | 87.2% | 2026-11-10 |
| 43 | GPT-6.1 Terra High | OpenAI | 86.8% | 2026-11-10 |
| 44 | Grok 4.20 (Reasoning) | xAI | 85.4% | 2026-11-10 |
| 45 | Qwen3.7 Max | Alibaba | 85.1% | 2026-11-10 |
| 46 | Muse Spark 1.3 High | Meta | 85.1% | 2026-11-10 |
| 47 | Muse Spark 1.1 High | Meta | 84.9% | 2026-11-10 |
| 48 | GPT-5.2 Extra High | OpenAI | 83.6% | 2026-11-10 |
| 49 | Hy4 High | Tencent | 83.5% | 2026-11-10 |
| 50 | Gemini 3 Flash Preview | 83.1% | 2026-11-10 | |
| 51 | Claude Sonnet 4.6 High | Anthropic | 80.9% | 2026-11-10 |
| 52 | Claude Sonnet 5.5 High | Anthropic | 80.5% | 2026-11-10 |
| 53 | Grok 4.6 Extra High | xAI | 80.0% | 2026-11-10 |
| 54 | Grok 4.5 High | xAI | 79.9% | 2026-11-10 |
| 55 | GPT-5.2 Pro | OpenAI | 79.3% | 2026-11-10 |
| 56 | Muse Spark 1.2 High | Meta | 79.2% | 2026-11-10 |
| 57 | Grok 4.6 High | xAI | 79.0% | 2026-11-10 |
| 58 | GPT-5.6 Terra High | OpenAI | 78.4% | 2026-11-10 |
| 59 | Muse Spark Open High | Meta | 78.4% | 2026-11-10 |
| 60 | GLM-5.1 | Zhipu AI | 77.7% | 2026-11-10 |
| 61 | Grok 4.7 High | xAI | 76.8% | 2026-11-10 |
| 62 | Step 5 Preview High | StepFun | 76.6% | 2026-11-10 |
| 63 | Claude Sonnet 4.6 Thinking | Anthropic | 76.4% | 2026-11-10 |
| 64 | Claude Opus 4.6 Thinking | Anthropic | 76.4% | 2026-11-10 |
| 65 | Claude Sonnet 5 High | Anthropic | 75.1% | 2026-11-10 |
| 66 | GLM-5 | Zhipu AI | 74.8% | 2026-11-10 |
| 67 | Qwen3.7 Plus | Alibaba | 74.8% | 2026-11-10 |
| 68 | GLM-5.2 High | Zhipu AI | 74.3% | 2026-11-10 |
| 69 | GLM-5.3 High | Zhipu AI | 74.2% | 2026-11-10 |
| 70 | Qwen3.6 Max Preview | Alibaba | 74.1% | 2026-11-10 |
| 71 | Gemma 4 31B IT Thinking | 70.6% | 2026-11-10 | |
| 72 | Kimi K2.5 Thinking | Moonshot AI | 69.9% | 2026-11-10 |
| 73 | GPT-5.6 Luna High | OpenAI | 69.4% | 2026-11-10 |
| 74 | GPT-6 Luna High | OpenAI | 68.7% | 2026-11-10 |
| 75 | Hy4 preview | Tencent | 68.2% | 2026-11-10 |
| 76 | GPT-5.2 High | OpenAI | 68.1% | 2026-11-10 |
| 77 | Claude Haiku 5.5 High | Anthropic | 65.7% | 2026-11-10 |
| 78 | MiniMax-M3 | MiniMax | 65.1% | 2026-11-10 |
| 79 | GPT-5.4 mini Extra High | OpenAI | 61.8% | 2026-11-10 |
| 80 | GPT-5.2 Medium | OpenAI | 60.6% | 2026-11-10 |
| 81 | Gemini 3.5 Flash Lite High | 60.4% | 2026-11-10 | |
| 82 | Qwen3.6 Plus | Alibaba | 60.3% | 2026-11-10 |
| 83 | Qwen3.5 397B-A17B | Alibaba | 58.9% | 2026-11-10 |
| 84 | Grok 4.3 | xAI | 55.2% | 2026-11-10 |
| 85 | Qwen3.8 27B | Alibaba | 54.5% | 2026-11-10 |
| 86 | GPT-5.2 Low | OpenAI | 54.2% | 2026-11-10 |
| 87 | Claude Opus 4.5 Thinking | Anthropic | 52.5% | 2026-11-10 |
| 88 | Qwen3.5 122B-A10B | Alibaba | 51.7% | 2026-11-10 |
| 89 | Claude Opus 4.5 Non-reasoning | Anthropic | 49.4% | 2026-11-10 |
| 90 | Qwen3.5 27B | Alibaba | 47.9% | 2026-11-10 |
| 91 | Claude Sonnet 4.6 Non-reasoning | Anthropic | 44.8% | 2026-11-10 |
| 92 | Qwen3.7 Flash | Alibaba | 43.8% | 2026-11-10 |
| 93 | Claude Opus 4.6 Non-reasoning | Anthropic | 43.2% | 2026-11-10 |
| 94 | Qwen3.6 35B-A3B | Alibaba | 41.6% | 2026-11-10 |
| 95 | Hy3 High | Tencent | 41.2% | 2026-11-10 |
| 96 | Step 3.7 Flash High | StepFun | 39.7% | 2026-11-10 |
| 97 | Claude Opus 4.7 High | Anthropic | 39.0% | 2026-11-10 |
| 98 | Claude Sonnet 4.5 Thinking | Anthropic | 37.3% | 2026-11-10 |
| 99 | DeepSeek V3.2 | DeepSeek | 36.7% | 2026-11-10 |
| 100 | Claude Sonnet 4.5 Non-reasoning | Anthropic | 35.8% | 2026-11-10 |
| 101 | MiMo-V2.5-Pro | Xiaomi | 34.4% | 2026-11-10 |
| 102 | Qwen3 Max | Alibaba | 30.1% | 2026-11-10 |
| 103 | Step 3.5 Flash | StepFun | 28.4% | 2026-11-10 |
| 104 | Seed 2.0 Pro | ByteDance | 28.4% | 2026-11-10 |
| 105 | MiniMax-M3.1-Flash High | MiniMax | 28.3% | 2026-11-10 |
| 106 | Mistral Large 4 High | Mistral | 27.4% | 2026-11-10 |
| 107 | MiMo-V2-Pro | Xiaomi | 25.8% | 2026-11-10 |
| 108 | MiniMax-M2.7 | MiniMax | 24.7% | 2026-11-10 |
| 109 | GPT-5.5 Non-reasoning | OpenAI | 22.0% | 2026-11-10 |
| 110 | GPT-5.4 Non-reasoning | OpenAI | 17.8% | 2026-11-10 |
| 111 | Hy3 preview | Tencent | 17.2% | 2026-11-10 |
| 112 | MiniMax-M2.5 | MiniMax | 16.8% | 2026-11-10 |
| 113 | Trinity Large Thinking | Arcee AI | 16.5% | 2026-11-10 |
| 114 | Gemma 4 31B IT | 15.7% | 2026-11-10 | |
| 115 | Nemotron 3 Super | NVIDIA | 15.4% | 2026-11-10 |
| 116 | MiniMax-M2 | MiniMax | 14.8% | 2026-11-10 |
| 117 | GPT-5.2 Non-reasoning | OpenAI | 14.5% | 2026-11-10 |
| 118 | Claude Haiku 4.5 (latest) | Anthropic | 14.3% | 2026-11-10 |
| 119 | Mistral Medium 3.5 High | Mistral | 12.9% | 2026-11-10 |
| 120 | Grok 4.1 Fast | xAI | 12.3% | 2026-11-10 |
| 121 | MiniMax-M2.1 | MiniMax | 11.2% | 2026-11-10 |
| 122 | Claude Opus 4.7 Non-reasoning | Anthropic | 10.8% | 2026-11-10 |
| 123 | Grok 4.20 (Non-Reasoning) | xAI | 8.6% | 2026-11-10 |
| 124 | Gemini 3.1 Flash Lite Preview | 8.2% | 2026-11-10 | |
| 125 | DeepSeek V3.2 Non-reasoning | DeepSeek | 8.2% | 2026-11-10 |
| 126 | Llama 4 Maverick | Meta | 8.0% | 2026-11-10 |
| 127 | Mistral Large 3 | Mistral | 7.5% | 2026-11-10 |
| 128 | Hy3 Non-reasoning | Tencent | 6.9% | 2026-11-10 |
| 129 | Mistral Medium 3.1 | Mistral | 6.5% | 2026-11-10 |