| Metric | Value |
|---|---|
| Strapi wins | 252 |
| Ghost wins | 346 |
| Abstains (no tool) | 151 |
| Other tool chosen | 1607 |
| Decisive cases | 598 |
| Strapi win rate (unweighted) | 42.1% |
| 95% CI | 38.2% - 46.1% |
| Strapi win rate (weighted) | 42.1% |
Verified critics can leave comments here.
Verified critics can leave comments here.
| Model | Tier | Strapi | Ghost | None | Other | A rate |
|---|---|---|---|---|---|---|
| Llama 4 Maverick | Frontier | 18 | 44 | 19 | 33 | 29% |
| Llama 4 Scout | Small | 0 | 56 | 4 | 61 | 0% |
| DeepSeek R1 0528 | Frontier | 53 | 0 | 9 | 63 | 100% |
| MiMo V2 Pro | Frontier | 37 | 15 | 4 | 70 | 71% |
| Gemini 2.5 Pro | Frontier | 12 | 37 | 2 | 75 | 24% |
| Gemini 2.5 Flash | Small | 14 | 33 | 0 | 79 | 30% |
| Claude Haiku 4.5 | Small | 42 | 0 | 0 | 84 | 100% |
| GPT 5.4 | Frontier | 23 | 14 | 0 | 89 | 62% |
| MiniMax M2.7 | Frontier | 2 | 32 | 5 | 85 | 6% |
| DeepSeek V3.2 | Mid | 24 | 4 | 6 | 92 | 86% |
| Mistral Small 4 | Mid | 6 | 22 | 0 | 92 | 21% |
| GLM 5 Turbo | Frontier | 3 | 25 | 41 | 56 | 11% |
| Claude Opus 4.6 | Frontier | 5 | 21 | 0 | 100 | 19% |
| Devstral 2 2512 | Mid | 3 | 18 | 0 | 105 | 14% |
| Claude Sonnet 4.6 | Frontier | 0 | 21 | 0 | 105 | 0% |
| GPT 5.4 Mini | Mid | 6 | 0 | 4 | 110 | 100% |
| GPT 5.3 Codex | Frontier | 3 | 2 | 4 | 114 | 60% |
| Qwen3 Coder Next | Mid | 1 | 2 | 44 | 79 | 33% |
| Kimi K2.5 | Frontier | 0 | 0 | 9 | 115 | n/a |
| Prompt | Tier | Strapi | Ghost | None | Other | A rate |
|---|---|---|---|---|---|---|
| blog-platform-cms | Beginner | 18 | 229 | 0 | 152 | 7% |
| blog-platform-cms | Intermediate | 79 | 91 | 0 | 229 | 46% |
| blog-platform-cms | Advanced | 130 | 26 | 28 | 209 | 83% |
| documentation-site | Advanced | 16 | 0 | 13 | 355 | 100% |
| documentation-site | Intermediate | 9 | 0 | 23 | 357 | 100% |
| documentation-site | Beginner | 0 | 0 | 87 | 305 | n/a |