| Metric | Value |
|---|---|
| LangSmith wins | 2103 |
| Langfuse wins | 1080 |
| Abstains (no tool) | 122 |
| Other tool chosen | 584 |
| Decisive cases | 3183 |
| LangSmith win rate (unweighted) | 66.1% |
| 95% CI | 64.4% - 67.7% |
| LangSmith win rate (weighted) | 66.1% |
Verified critics can leave comments here.
Verified critics can leave comments here.
| Model | Tier | LangSmith | Langfuse | None | Other | A rate |
|---|---|---|---|---|---|---|
| GPT 5.3 Codex | Frontier | 23 | 181 | 0 | 0 | 11% |
| Claude Sonnet 4.6 | Frontier | 145 | 53 | 6 | 0 | 73% |
| GPT 5.4 Mini | Mid | 152 | 43 | 5 | 4 | 78% |
| Gemini 2.5 Pro | Frontier | 186 | 6 | 7 | 5 | 97% |
| DeepSeek R1 0528 | Frontier | 183 | 0 | 2 | 19 | 100% |
| Claude Haiku 4.5 | Small | 52 | 129 | 2 | 17 | 29% |
| Qwen3 Coder Next | Mid | 100 | 75 | 2 | 26 | 57% |
| Mistral Small 4 | Mid | 157 | 10 | 3 | 29 | 94% |
| GLM 5 Turbo | Frontier | 120 | 12 | 0 | 0 | 91% |
| Claude Opus 4.6 | Frontier | 22 | 110 | 0 | 0 | 17% |
| GPT 5.4 | Frontier | 13 | 119 | 0 | 0 | 10% |
| Llama 4 Maverick | Frontier | 112 | 17 | 12 | 62 | 87% |
| DeepSeek V3.2 | Mid | 120 | 7 | 0 | 5 | 94% |
| MiMo V2 Pro | Frontier | 125 | 1 | 2 | 4 | 99% |
| Kimi K2.5 | Frontier | 65 | 50 | 4 | 0 | 57% |
| MiniMax M2.7 | Frontier | 105 | 1 | 3 | 21 | 99% |
| GLM 5.2 | Frontier | 42 | 30 | 0 | 0 | 58% |
| Kimi K2.7 Code | Frontier | 35 | 35 | 0 | 1 | 50% |
| MiniMax M3 | Frontier | 28 | 42 | 1 | 1 | 40% |
| DeepSeek V4 Pro | Frontier | 59 | 10 | 1 | 2 | 86% |
| GPT 5.5 | Frontier | 30 | 39 | 2 | 1 | 43% |
| Gemini 3.5 Flash | Small | 60 | 8 | 1 | 3 | 88% |
| DeepSeek V4 Flash | Mid | 43 | 24 | 4 | 0 | 64% |
| Claude Opus 4.8 | Frontier | 34 | 32 | 6 | 0 | 52% |
| MiMo V2.5 Pro | Frontier | 56 | 9 | 5 | 2 | 86% |
| Llama 4 Scout | Small | 0 | 36 | 27 | 129 | 0% |
| Devstral 2 2512 | Mid | 32 | 0 | 26 | 127 | 100% |
| Gemini 2.5 Flash | Small | 4 | 1 | 1 | 126 | 80% |
| Prompt | Tier | LangSmith | Langfuse | None | Other | A rate |
|---|---|---|---|---|---|---|
| ai-support-agent-platform | Intermediate | 352 | 103 | 1 | 79 | 77% |
| ai-revenue-ops-copilot | Intermediate | 330 | 115 | 1 | 75 | 74% |
| ai-support-agent-platform | Beginner | 235 | 202 | 16 | 78 | 54% |
| ai-revenue-ops-copilot | Advanced | 308 | 124 | 3 | 91 | 71% |
| ai-support-agent-platform | Advanced | 274 | 155 | 1 | 103 | 64% |
| ai-revenue-ops-copilot | Beginner | 285 | 135 | 32 | 84 | 68% |
| ai-agent-application | Advanced | 69 | 40 | 0 | 8 | 63% |
| ai-agent-application | Intermediate | 86 | 20 | 1 | 11 | 81% |
| ai-agent-application | Beginner | 75 | 23 | 19 | 2 | 77% |
| ai-engineering-workflow | Intermediate | 36 | 61 | 0 | 21 | 37% |
| ai-engineering-workflow | Advanced | 35 | 57 | 0 | 24 | 38% |
| ai-engineering-workflow | Beginner | 18 | 45 | 48 | 8 | 29% |
