| Metric | Value |
|---|---|
| Sanity wins | 523 |
| Contentful wins | 368 |
| Abstains (no tool) | 151 |
| Other tool chosen | 1314 |
| Decisive cases | 891 |
| Sanity win rate (unweighted) | 58.7% |
| 95% CI | 55.4% - 61.9% |
| Sanity win rate (weighted) | 58.7% |
Verified critics can leave comments here.
Verified critics can leave comments here.
| Model | Tier | Sanity | Contentful | None | Other | A rate |
|---|---|---|---|---|---|---|
| GPT 5.4 Mini | Mid | 67 | 25 | 4 | 24 | 73% |
| Kimi K2.5 | Frontier | 87 | 1 | 9 | 27 | 99% |
| GPT 5.3 Codex | Frontier | 58 | 29 | 4 | 32 | 67% |
| Qwen3 Coder Next | Mid | 44 | 26 | 44 | 12 | 63% |
| Mistral Small 4 | Mid | 35 | 34 | 0 | 51 | 51% |
| DeepSeek V3.2 | Mid | 43 | 19 | 6 | 58 | 69% |
| Claude Opus 4.6 | Frontier | 42 | 16 | 0 | 68 | 72% |
| GPT 5.4 | Frontier | 40 | 17 | 0 | 69 | 70% |
| Gemini 2.5 Flash | Small | 0 | 56 | 0 | 70 | 0% |
| GLM 5 Turbo | Frontier | 34 | 9 | 41 | 41 | 79% |
| Gemini 2.5 Pro | Frontier | 18 | 19 | 2 | 87 | 49% |
| MiniMax M2.7 | Frontier | 15 | 22 | 5 | 82 | 41% |
| DeepSeek R1 0528 | Frontier | 13 | 22 | 9 | 81 | 37% |
| MiMo V2 Pro | Frontier | 9 | 23 | 4 | 90 | 28% |
| Claude Haiku 4.5 | Small | 0 | 26 | 0 | 100 | 0% |
| Devstral 2 2512 | Mid | 0 | 21 | 0 | 105 | 0% |
| Claude Sonnet 4.6 | Frontier | 18 | 0 | 0 | 108 | 100% |
| Llama 4 Scout | Small | 0 | 2 | 4 | 115 | 0% |
| Llama 4 Maverick | Frontier | 0 | 1 | 19 | 94 | 0% |
| Prompt | Tier | Sanity | Contentful | None | Other | A rate |
|---|---|---|---|---|---|---|
| documentation-site | Advanced | 113 | 171 | 13 | 87 | 40% |
| blog-platform-cms | Intermediate | 164 | 14 | 0 | 221 | 92% |
| documentation-site | Intermediate | 68 | 92 | 23 | 206 | 43% |
| blog-platform-cms | Advanced | 81 | 50 | 28 | 234 | 62% |
| blog-platform-cms | Beginner | 88 | 29 | 0 | 282 | 75% |
| documentation-site | Beginner | 9 | 12 | 87 | 284 | 43% |
