| Metric | Value |
|---|---|
| Ragas wins | 215 |
| LangSmith wins | 1338 |
| Abstains (no tool) | 237 |
| Other tool chosen | 2738 |
| Decisive cases | 1553 |
| Ragas win rate (unweighted) | 13.8% |
| 95% CI | 12.2% - 15.7% |
| Ragas win rate (weighted) | 13.8% |
Verified critics can leave comments here.
Verified critics can leave comments here.
| Model range | Tier | Ragas | LangSmith | None | Other | A rate |
|---|---|---|---|---|---|---|
| DeepSeek Flash 3.2–4.1 | Mid | 9 | 109 | 26 | 92 | 8% |
| DeepSeek Pro 4–4 0813 | Frontier | 7 | 25 | 2 | 74 | 22% |
| DeepSeek R1 | Frontier | 0 | 200 | 12 | 28 | 0% |
| Devstral | Mid | 4 | 10 | 19 | 168 | 29% |
| Gemini Flash 2.5–3.8 | Small | 2 | 3 | 12 | 218 | 40% |
| Gemini Pro 2.5–3.1 Preview | Frontier | 5 | 159 | 17 | 57 | 3% |
| GLM 5 Turbo–5.3 | Frontier | 25 | 23 | 19 | 173 | 52% |
| GPT 5.4–6 Astra | Frontier | 3 | 26 | 5 | 206 | 10% |
| GPT Coding 5.3 Codex–6.1 Sol | Frontier | 0 | 17 | 2 | 221 | 0% |
| GPT Mini 5.4 Mini–6 Luna | Mid | 17 | 140 | 9 | 71 | 11% |
| Haiku | Small | 1 | 20 | 4 | 206 | 5% |
| Kimi 2.5–3 | Frontier | 10 | 1 | 5 | 210 | 91% |
| Llama Maverick | Frontier | 0 | 44 | 18 | 169 | 0% |
| Llama Scout | Small | 5 | 0 | 29 | 180 | 100% |
| MiMo Pro 2–2.6 | Frontier | 32 | 81 | 17 | 109 | 28% |
| MiniMax M 2.7–3 | Frontier | 31 | 64 | 12 | 129 | 33% |
| Mistral Small | Mid | 19 | 146 | 5 | 56 | 12% |
| Opus 4.6–5.5 | Frontier | 31 | 34 | 9 | 165 | 48% |
| Qwen Coder | Mid | 0 | 158 | 4 | 69 | 0% |
| Sonnet 4.6–5.5 | Frontier | 14 | 78 | 11 | 137 | 15% |
| Prompt | Tier | Ragas | LangSmith | None | Other | A rate |
|---|---|---|---|---|---|---|
| ai-support-agent-platform | Intermediate | 9 | 276 | 6 | 289 | 3% |
| ai-revenue-ops-copilot | Intermediate | 2 | 243 | 5 | 330 | 1% |
| ai-support-agent-platform | Advanced | 39 | 184 | 5 | 359 | 17% |
| ai-revenue-ops-copilot | Beginner | 17 | 162 | 13 | 392 | 9% |
| ai-revenue-ops-copilot | Advanced | 11 | 165 | 3 | 395 | 6% |
| ai-support-agent-platform | Beginner | 39 | 126 | 80 | 339 | 24% |
| ai-agent-application | Beginner | 41 | 46 | 40 | 51 | 47% |
| ai-agent-application | Intermediate | 36 | 39 | 1 | 99 | 48% |
| ai-agent-application | Advanced | 19 | 35 | 0 | 119 | 35% |
| ai-engineering-workflow | Intermediate | 1 | 22 | 1 | 149 | 4% |
| ai-engineering-workflow | Advanced | 0 | 21 | 1 | 146 | 0% |
| ai-engineering-workflow | Beginner | 1 | 19 | 82 | 70 | 5% |
