| Metric | Value |
|---|---|
| TypeScript wins | 191 |
| Python wins | 525 |
| Abstains (no tool) | 23 |
| Other tool chosen | 49 |
| Decisive cases | 716 |
| TypeScript win rate (unweighted) | 26.7% |
| 95% CI | 23.6% - 30.0% |
| TypeScript win rate (weighted) | 26.7% |
Verified critics can leave comments here.
Verified critics can leave comments here.
| Model | Tier | TypeScript | Python | None | Other | A rate |
|---|---|---|---|---|---|---|
| GLM 5.2 | Frontier | 20 | 22 | 0 | 0 | 48% |
| Claude Opus 4.8 | Frontier | 11 | 31 | 0 | 0 | 26% |
| Gemini 3.5 Flash | Small | 13 | 28 | 0 | 1 | 32% |
| Claude Sonnet 4.6 | Frontier | 11 | 30 | 0 | 1 | 27% |
| GPT 5.3 Codex | Frontier | 11 | 30 | 1 | 0 | 27% |
| GPT 5.4 Mini | Mid | 19 | 21 | 0 | 2 | 48% |
| MiniMax M3 | Frontier | 17 | 23 | 0 | 0 | 43% |
| MiMo V2.5 Pro | Frontier | 5 | 35 | 0 | 0 | 13% |
| GPT 5.5 | Frontier | 14 | 25 | 0 | 3 | 36% |
| DeepSeek V4 Pro | Frontier | 11 | 27 | 2 | 2 | 29% |
| Gemini 2.5 Pro | Frontier | 10 | 27 | 1 | 4 | 27% |
| Kimi K2.7 Code | Frontier | 18 | 18 | 4 | 2 | 50% |
| Claude Haiku 4.5 | Small | 1 | 34 | 0 | 7 | 3% |
| Llama 4 Scout | Small | 0 | 34 | 0 | 0 | 0% |
| DeepSeek V4 Flash | Mid | 3 | 29 | 0 | 8 | 9% |
| DeepSeek R1 0528 | Frontier | 2 | 30 | 5 | 4 | 6% |
| Qwen3 Coder Next | Mid | 11 | 20 | 4 | 0 | 35% |
| Devstral 2 2512 | Mid | 8 | 22 | 0 | 0 | 27% |
| Llama 4 Maverick | Frontier | 5 | 19 | 6 | 7 | 21% |
| Mistral Small 4 | Mid | 1 | 20 | 0 | 8 | 5% |
| Prompt | Tier | TypeScript | Python | None | Other | A rate |
|---|---|---|---|---|---|---|
| ai-agent-application | Beginner | 1 | 134 | 2 | 1 | 1% |
| ai-agent-application | Intermediate | 1 | 130 | 7 | 0 | 1% |
| ai-agent-application | Advanced | 1 | 125 | 10 | 0 | 1% |
| full-stack-web-app | Intermediate | 75 | 40 | 0 | 4 | 65% |
| full-stack-web-app | Beginner | 32 | 77 | 3 | 9 | 29% |
| full-stack-web-app | Advanced | 81 | 19 | 1 | 35 | 81% |
