Leading: Browserbase (95.8%)
| Metric | Value |
|---|---|
| Browser Use wins | 3 |
| Browserbase wins | 69 |
| Abstains (no tool) | 100 |
| Other tool chosen | 163 |
| Decisive cases | 72 |
| Browser Use win rate (unweighted) | 4.2% |
| 95% CI | 1.4% - 11.5% |
| Browser Use win rate (weighted) | 4.2% |
Verified critics can leave comments here.
Verified critics can leave comments here.
| Model | Tier | Browser Use | Browserbase | None | Other | A rate |
|---|---|---|---|---|---|---|
| MiniMax M3 | Frontier | 2 | 11 | 2 | 2 | 15% |
| Gemini 3.5 Flash | Small | 0 | 12 | 4 | 2 | 0% |
| GLM 5.2 | Frontier | 1 | 10 | 6 | 1 | 9% |
| GPT 5.5 | Frontier | 0 | 11 | 5 | 2 | 0% |
| Kimi K2.7 Code | Frontier | 0 | 7 | 6 | 5 | 0% |
| MiMo V2.5 Pro | Frontier | 0 | 5 | 6 | 7 | 0% |
| Claude Haiku 4.5 | Small | 0 | 4 | 6 | 8 | 0% |
| Claude Opus 4.8 | Frontier | 0 | 4 | 6 | 8 | 0% |
| DeepSeek V4 Pro | Frontier | 0 | 3 | 5 | 10 | 0% |
| Gemini 2.5 Pro | Frontier | 0 | 1 | 5 | 4 | 0% |
| GPT 5.4 Mini | Mid | 0 | 1 | 1 | 16 | 0% |
| Claude Sonnet 4.6 | Frontier | 0 | 0 | 6 | 12 | n/a |
| DeepSeek R1 0528 | Frontier | 0 | 0 | 6 | 12 | n/a |
| DeepSeek V4 Flash | Mid | 0 | 0 | 4 | 14 | n/a |
| Devstral 2 2512 | Mid | 0 | 0 | 5 | 9 | n/a |
| GPT 5.3 Codex | Frontier | 0 | 0 | 5 | 13 | n/a |
| Llama 4 Maverick | Frontier | 0 | 0 | 6 | 12 | n/a |
| Llama 4 Scout | Small | 0 | 0 | 6 | 6 | n/a |
| Mistral Small 4 | Mid | 0 | 0 | 4 | 14 | n/a |
| Qwen3 Coder Next | Mid | 0 | 0 | 6 | 6 | n/a |
| Prompt | Tier | Browser Use | Browserbase | None | Other | A rate |
|---|---|---|---|---|---|---|
| ai-agent-application | Intermediate | 2 | 33 | 1 | 75 | 6% |
| ai-agent-application | Advanced | 1 | 32 | 0 | 73 | 3% |
| ai-agent-application | Beginner | 0 | 4 | 99 | 15 | 0% |
