| Metric | Value |
|---|---|
| Go wins | 28 |
| TypeScript wins | 191 |
| Abstains (no tool) | 23 |
| Other tool chosen | 546 |
| Decisive cases | 219 |
| Go win rate (unweighted) | 12.8% |
| 95% CI | 9.0% - 17.9% |
| Go win rate (weighted) | 12.8% |
Verified critics can leave comments here.
Verified critics can leave comments here.
| Model | Tier | Go | TypeScript | None | Other | A rate |
|---|---|---|---|---|---|---|
| GPT 5.4 Mini | Mid | 2 | 19 | 0 | 21 | 10% |
| GLM 5.2 | Frontier | 0 | 20 | 0 | 22 | 0% |
| Kimi K2.7 Code | Frontier | 1 | 18 | 4 | 19 | 5% |
| MiniMax M3 | Frontier | 0 | 17 | 0 | 23 | 0% |
| Gemini 3.5 Flash | Small | 1 | 13 | 0 | 28 | 7% |
| GPT 5.5 | Frontier | 0 | 14 | 0 | 28 | 0% |
| DeepSeek V4 Pro | Frontier | 1 | 11 | 2 | 28 | 8% |
| DeepSeek V4 Flash | Mid | 8 | 3 | 0 | 29 | 73% |
| Claude Opus 4.8 | Frontier | 0 | 11 | 0 | 31 | 0% |
| Claude Sonnet 4.6 | Frontier | 0 | 11 | 0 | 31 | 0% |
| GPT 5.3 Codex | Frontier | 0 | 11 | 1 | 30 | 0% |
| Qwen3 Coder Next | Mid | 0 | 11 | 4 | 20 | 0% |
| Gemini 2.5 Pro | Frontier | 0 | 10 | 1 | 31 | 0% |
| Claude Haiku 4.5 | Small | 7 | 1 | 0 | 34 | 88% |
| Mistral Small 4 | Mid | 7 | 1 | 0 | 21 | 88% |
| Devstral 2 2512 | Mid | 0 | 8 | 0 | 22 | 0% |
| Llama 4 Maverick | Frontier | 0 | 5 | 6 | 26 | 0% |
| MiMo V2.5 Pro | Frontier | 0 | 5 | 0 | 35 | 0% |
| DeepSeek R1 0528 | Frontier | 1 | 2 | 5 | 33 | 33% |
| Llama 4 Scout | Small | 0 | 0 | 0 | 34 | n/a |
| Prompt | Tier | Go | TypeScript | None | Other | A rate |
|---|---|---|---|---|---|---|
| full-stack-web-app | Advanced | 26 | 81 | 1 | 28 | 24% |
| full-stack-web-app | Intermediate | 2 | 75 | 0 | 42 | 3% |
| full-stack-web-app | Beginner | 0 | 32 | 3 | 86 | 0% |
| ai-agent-application | Advanced | 0 | 1 | 10 | 125 | 0% |
| ai-agent-application | Beginner | 0 | 1 | 2 | 135 | 0% |
| ai-agent-application | Intermediate | 0 | 1 | 7 | 130 | 0% |
