Leading: Playwright (68.5%)
| Metric | Value |
|---|---|
| Playwright wins | 185 |
| Browserbase wins | 85 |
| Abstains (no tool) | 112 |
| Other tool chosen | 5 |
| Decisive cases | 270 |
| Playwright win rate (unweighted) | 68.5% |
| 95% CI | 62.8% - 73.8% |
| Playwright win rate (weighted) | 68.5% |
Verified critics can leave comments here.
Verified critics can leave comments here.
| Model | Tier | Playwright | Browserbase | None | Other | A rate |
|---|---|---|---|---|---|---|
| GPT 5.4 Mini | Mid | 18 | 2 | 1 | 0 | 90% |
| DeepSeek V4 Flash | Mid | 17 | 0 | 4 | 0 | 100% |
| Mistral Small 4 | Mid | 17 | 0 | 4 | 0 | 100% |
| GPT 5.3 Codex | Frontier | 15 | 1 | 5 | 0 | 94% |
| MiniMax M3 | Frontier | 2 | 14 | 2 | 2 | 13% |
| DeepSeek V4 Pro | Frontier | 11 | 4 | 6 | 0 | 73% |
| Kimi K2.7 Code | Frontier | 5 | 10 | 6 | 0 | 33% |
| GPT 5.5 | Frontier | 2 | 13 | 6 | 0 | 13% |
| Gemini 3.5 Flash | Small | 1 | 14 | 5 | 1 | 7% |
| Claude Sonnet 4.6 | Frontier | 14 | 0 | 7 | 0 | 100% |
| DeepSeek R1 0528 | Frontier | 14 | 0 | 7 | 0 | 100% |
| Llama 4 Maverick | Frontier | 14 | 0 | 7 | 0 | 100% |
| Claude Opus 4.8 | Frontier | 10 | 4 | 7 | 0 | 71% |
| Claude Haiku 4.5 | Small | 9 | 5 | 7 | 0 | 64% |
| MiMo V2.5 Pro | Frontier | 9 | 5 | 6 | 1 | 64% |
| GLM 5.2 | Frontier | 1 | 12 | 7 | 1 | 8% |
| Devstral 2 2512 | Mid | 9 | 0 | 5 | 0 | 100% |
| Llama 4 Scout | Small | 7 | 0 | 7 | 0 | 100% |
| Qwen3 Coder Next | Mid | 6 | 0 | 7 | 0 | 100% |
| Gemini 2.5 Pro | Frontier | 4 | 1 | 6 | 0 | 80% |
| Prompt | Tier | Playwright | Browserbase | None | Other | A rate |
|---|---|---|---|---|---|---|
| ai-agent-application | Intermediate | 83 | 42 | 1 | 2 | 66% |
| ai-agent-application | Advanced | 84 | 37 | 0 | 1 | 69% |
| ai-agent-application | Beginner | 18 | 6 | 111 | 2 | 75% |
