| Metric | Value |
|---|---|
| Python wins | 456 |
| TypeScript wins | 162 |
| Abstains (no tool) | 20 |
| Other tool chosen | 43 |
| Decisive cases | 618 |
| Python win rate (unweighted) | 73.8% |
| 95% CI | 70.2% - 77.1% |
| Python win rate (weighted) | 73.8% |
Verified critics can leave comments here.
Verified critics can leave comments here.
| Model | Tier | Python | TypeScript | None | Other | A rate |
|---|---|---|---|---|---|---|
| Claude Opus 4.8 | Frontier | 27 | 9 | 0 | 0 | 75% |
| GLM 5.2 | Frontier | 19 | 17 | 0 | 0 | 53% |
| Claude Sonnet 4.6 | Frontier | 26 | 9 | 0 | 1 | 74% |
| GPT 5.3 Codex | Frontier | 25 | 10 | 1 | 0 | 71% |
| Gemini 3.5 Flash | Small | 24 | 11 | 0 | 1 | 69% |
| MiniMax M3 | Frontier | 20 | 15 | 0 | 0 | 57% |
| MiMo V2.5 Pro | Frontier | 30 | 4 | 0 | 0 | 88% |
| GPT 5.4 Mini | Mid | 17 | 17 | 0 | 2 | 50% |
| GPT 5.5 | Frontier | 22 | 11 | 0 | 3 | 67% |
| Gemini 2.5 Pro | Frontier | 24 | 8 | 0 | 4 | 75% |
| DeepSeek V4 Pro | Frontier | 23 | 9 | 2 | 2 | 72% |
| Kimi K2.7 Code | Frontier | 16 | 15 | 3 | 2 | 52% |
| Llama 4 Scout | Small | 30 | 0 | 0 | 0 | 100% |
| Claude Haiku 4.5 | Small | 29 | 1 | 0 | 6 | 97% |
| Devstral 2 2512 | Mid | 22 | 8 | 0 | 0 | 73% |
| DeepSeek V4 Flash | Mid | 25 | 3 | 0 | 6 | 89% |
| DeepSeek R1 0528 | Frontier | 26 | 1 | 5 | 3 | 96% |
| Qwen3 Coder Next | Mid | 17 | 9 | 4 | 0 | 65% |
| Llama 4 Maverick | Frontier | 17 | 4 | 5 | 6 | 81% |
| Mistral Small 4 | Mid | 17 | 1 | 0 | 7 | 94% |
| Prompt | Tier | Python | TypeScript | None | Other | A rate |
|---|---|---|---|---|---|---|
| ai-agent-application | Beginner | 115 | 1 | 2 | 1 | 99% |
| ai-agent-application | Intermediate | 112 | 1 | 6 | 0 | 99% |
| ai-agent-application | Advanced | 107 | 1 | 9 | 0 | 99% |
| full-stack-web-app | Intermediate | 37 | 63 | 0 | 3 | 37% |
| full-stack-web-app | Beginner | 68 | 26 | 3 | 9 | 72% |
| full-stack-web-app | Advanced | 17 | 70 | 0 | 30 | 20% |
