| Metric | Value |
|---|---|
| LangSmith wins | 902 |
| Braintrust wins | 806 |
| Abstains (no tool) | 105 |
| Other tool chosen | 862 |
| Decisive cases | 1708 |
| LangSmith win rate (unweighted) | 52.8% |
| 95% CI | 50.4% - 55.2% |
| LangSmith win rate (weighted) | 52.8% |
Verified critics can leave comments here.
Verified critics can leave comments here.
| Model | Tier | LangSmith | Braintrust | None | Other | A rate |
|---|---|---|---|---|---|---|
| GPT 5.3 Codex | Frontier | 8 | 136 | 0 | 0 | 6% |
| Claude Haiku 4.5 | Small | 14 | 113 | 1 | 9 | 11% |
| Claude Sonnet 4.6 | Frontier | 58 | 65 | 1 | 20 | 47% |
| DeepSeek R1 0528 | Frontier | 119 | 0 | 7 | 18 | 100% |
| Gemini 2.5 Pro | Frontier | 116 | 1 | 9 | 18 | 99% |
| Claude Opus 4.6 | Frontier | 1 | 113 | 0 | 18 | 1% |
| GPT 5.4 | Frontier | 18 | 92 | 0 | 22 | 16% |
| Kimi K2.5 | Frontier | 0 | 109 | 3 | 7 | 0% |
| GPT 5.4 Mini | Mid | 102 | 4 | 3 | 34 | 96% |
| GLM 5 Turbo | Frontier | 18 | 84 | 19 | 11 | 18% |
| Mistral Small 4 | Mid | 101 | 0 | 2 | 33 | 100% |
| Qwen3 Coder Next | Mid | 95 | 0 | 3 | 43 | 100% |
| MiniMax M2.7 | Frontier | 62 | 31 | 5 | 31 | 67% |
| DeepSeek V3.2 | Mid | 80 | 0 | 22 | 26 | 100% |
| MiMo V2 Pro | Frontier | 62 | 1 | 8 | 61 | 98% |
| Llama 4 Maverick | Frontier | 22 | 0 | 2 | 113 | 100% |
| GPT 5.5 | Frontier | 0 | 12 | 0 | 0 | 0% |
| MiniMax M3 | Frontier | 0 | 10 | 1 | 1 | 0% |
| DeepSeek V4 Pro | Frontier | 5 | 4 | 1 | 2 | 56% |
| Gemini 3.5 Flash | Small | 0 | 9 | 1 | 2 | 0% |
| GLM 5.2 | Frontier | 0 | 9 | 0 | 3 | 0% |
| MiMo V2.5 Pro | Frontier | 3 | 5 | 0 | 4 | 38% |
| Devstral 2 2512 | Mid | 6 | 0 | 6 | 123 | 100% |
| Claude Opus 4.8 | Frontier | 5 | 1 | 1 | 5 | 83% |
| Kimi K2.7 Code | Frontier | 0 | 6 | 1 | 5 | 0% |
| DeepSeek V4 Flash | Mid | 4 | 1 | 1 | 6 | 80% |
| Gemini 2.5 Flash | Small | 3 | 0 | 1 | 123 | 100% |
| Llama 4 Scout | Small | 0 | 0 | 7 | 124 | n/a |
| Prompt | Tier | LangSmith | Braintrust | None | Other | A rate |
|---|---|---|---|---|---|---|
| ai-revenue-ops-copilot | Intermediate | 189 | 130 | 4 | 101 | 59% |
| ai-support-agent-platform | Intermediate | 207 | 92 | 5 | 124 | 69% |
| ai-revenue-ops-copilot | Advanced | 128 | 164 | 2 | 125 | 44% |
| ai-revenue-ops-copilot | Beginner | 126 | 151 | 10 | 143 | 45% |
| ai-support-agent-platform | Advanced | 142 | 130 | 5 | 153 | 52% |
| ai-support-agent-platform | Beginner | 87 | 100 | 66 | 178 | 47% |
| ai-engineering-workflow | Advanced | 2 | 12 | 0 | 4 | 14% |
| ai-agent-application | Beginner | 8 | 3 | 5 | 4 | 73% |
| ai-agent-application | Intermediate | 5 | 6 | 0 | 9 | 45% |
| ai-agent-application | Advanced | 3 | 7 | 0 | 8 | 30% |
| ai-engineering-workflow | Intermediate | 2 | 8 | 0 | 7 | 20% |
| ai-engineering-workflow | Beginner | 3 | 3 | 8 | 6 | 50% |
