| Metric | Value |
|---|---|
| DeepEval wins | 162 |
| Ragas wins | 191 |
| Abstains (no tool) | 194 |
| Other tool chosen | 3390 |
| Decisive cases | 353 |
| DeepEval win rate (unweighted) | 45.9% |
| 95% CI | 40.8% - 51.1% |
| DeepEval win rate (weighted) | 45.9% |
Verified critics can leave comments here.
Verified critics can leave comments here.
| Model | Tier | DeepEval | Ragas | None | Other | A rate |
|---|---|---|---|---|---|---|
| Llama 4 Maverick | Frontier | 111 | 0 | 14 | 70 | 100% |
| DeepSeek V4 Flash | Mid | 28 | 6 | 4 | 45 | 82% |
| MiMo V2 Pro | Frontier | 0 | 23 | 8 | 95 | 0% |
| MiniMax M2.7 | Frontier | 0 | 23 | 5 | 95 | 0% |
| Claude Opus 4.6 | Frontier | 0 | 17 | 0 | 109 | 0% |
| GLM 5.2 | Frontier | 0 | 16 | 0 | 68 | 0% |
| GPT 5.4 Mini | Mid | 0 | 16 | 8 | 183 | 0% |
| Mistral Small 4 | Mid | 0 | 16 | 3 | 177 | 0% |
| Claude Sonnet 4.6 | Frontier | 0 | 13 | 7 | 190 | 0% |
| DeepSeek V4 Pro | Frontier | 5 | 7 | 2 | 70 | 42% |
| MiMo V2.5 Pro | Frontier | 4 | 8 | 9 | 62 | 33% |
| GPT 5.4 | Frontier | 9 | 1 | 0 | 116 | 90% |
| Claude Opus 4.8 | Frontier | 0 | 9 | 7 | 67 | 0% |
| Kimi K2.7 Code | Frontier | 0 | 7 | 1 | 75 | 0% |
| GLM 5 Turbo | Frontier | 0 | 6 | 19 | 101 | 0% |
| Kimi K2.5 | Frontier | 4 | 1 | 3 | 106 | 80% |
| Llama 4 Scout | Small | 0 | 5 | 23 | 161 | 0% |
| MiniMax M3 | Frontier | 0 | 5 | 6 | 73 | 0% |
| Devstral 2 2512 | Mid | 0 | 4 | 14 | 154 | 0% |
| DeepSeek V3.2 | Mid | 1 | 2 | 22 | 98 | 33% |
| Gemini 2.5 Pro | Frontier | 0 | 3 | 9 | 196 | 0% |
| GPT 5.5 | Frontier | 0 | 2 | 3 | 79 | 0% |
| Claude Haiku 4.5 | Small | 0 | 1 | 4 | 195 | 0% |
| DeepSeek R1 0528 | Frontier | 0 | 0 | 10 | 200 | n/a |
| Gemini 2.5 Flash | Small | 0 | 0 | 1 | 120 | n/a |
| Gemini 3.5 Flash | Small | 0 | 0 | 8 | 76 | n/a |
| GPT 5.3 Codex | Frontier | 0 | 0 | 0 | 210 | n/a |
| Qwen3 Coder Next | Mid | 0 | 0 | 4 | 199 | n/a |
| Prompt | Tier | DeepEval | Ragas | None | Other | A rate |
|---|---|---|---|---|---|---|
| ai-support-agent-platform | Advanced | 39 | 37 | 5 | 446 | 51% |
| ai-support-agent-platform | Beginner | 26 | 36 | 75 | 391 | 42% |
| ai-agent-application | Intermediate | 6 | 29 | 1 | 96 | 17% |
| ai-agent-application | Beginner | 1 | 31 | 28 | 78 | 3% |
| ai-support-agent-platform | Intermediate | 22 | 9 | 5 | 486 | 71% |
| ai-agent-application | Advanced | 12 | 18 | 0 | 101 | 40% |
| ai-revenue-ops-copilot | Beginner | 9 | 16 | 11 | 489 | 36% |
| ai-revenue-ops-copilot | Advanced | 13 | 11 | 2 | 490 | 54% |
| ai-engineering-workflow | Advanced | 11 | 0 | 1 | 118 | 100% |
| ai-revenue-ops-copilot | Intermediate | 9 | 2 | 4 | 508 | 82% |
| ai-engineering-workflow | Intermediate | 9 | 1 | 1 | 120 | 90% |
| ai-engineering-workflow | Beginner | 5 | 1 | 61 | 67 | 83% |
