| Metric | Value |
|---|---|
| Codex CLI wins | 2 |
| OpenHands wins | 2 |
| Abstains (no tool) | 284 |
| Other tool chosen | 712 |
| Decisive cases | 4 |
| Codex CLI win rate (unweighted) | 50.0% |
| 95% CI | 15.0% - 85.0% |
| Codex CLI win rate (weighted) | 50.0% |
Verified critics can leave comments here.
Verified critics can leave comments here.
| Model | Tier | Codex CLI | OpenHands | None | Other | A rate |
|---|---|---|---|---|---|---|
| GPT 5.4 | Frontier | 1 | 1 | 10 | 39 | 50% |
| GPT 5.4 Mini | Mid | 1 | 1 | 4 | 48 | 50% |
| Claude Haiku 4.5 | Small | 0 | 0 | 26 | 25 | n/a |
| Claude Opus 4.6 | Frontier | 0 | 0 | 8 | 46 | n/a |
| Claude Sonnet 4.6 | Frontier | 0 | 0 | 22 | 32 | n/a |
| DeepSeek R1 0528 | Frontier | 0 | 0 | 21 | 33 | n/a |
| DeepSeek V3.2 | Mid | 0 | 0 | 21 | 33 | n/a |
| Devstral 2 2512 | Mid | 0 | 0 | 19 | 32 | n/a |
| Gemini 2.5 Flash | Small | 0 | 0 | 25 | 29 | n/a |
| Gemini 2.5 Pro | Frontier | 0 | 0 | 6 | 48 | n/a |
| GLM 5 Turbo | Frontier | 0 | 0 | 0 | 54 | n/a |
| GPT 5.3 Codex | Frontier | 0 | 0 | 10 | 44 | n/a |
| Kimi K2.5 | Frontier | 0 | 0 | 21 | 27 | n/a |
| Llama 4 Maverick | Frontier | 0 | 0 | 22 | 32 | n/a |
| Llama 4 Scout | Small | 0 | 0 | 18 | 35 | n/a |
| MiMo V2 Pro | Frontier | 0 | 0 | 12 | 42 | n/a |
| MiniMax M2.7 | Frontier | 0 | 0 | 11 | 42 | n/a |
| Mistral Small 4 | Mid | 0 | 0 | 22 | 24 | n/a |
| Qwen3 Coder Next | Mid | 0 | 0 | 6 | 47 | n/a |
| Prompt | Tier | Codex CLI | OpenHands | None | Other | A rate |
|---|---|---|---|---|---|---|
| ai-support-agent-platform | Advanced | 2 | 2 | 0 | 165 | 50% |
| ai-revenue-ops-copilot | Intermediate | 0 | 0 | 1 | 163 | n/a |
| ai-revenue-ops-copilot | Beginner | 0 | 0 | 90 | 79 | n/a |
| ai-revenue-ops-copilot | Advanced | 0 | 0 | 2 | 156 | n/a |
| ai-support-agent-platform | Intermediate | 0 | 0 | 81 | 89 | n/a |
| ai-support-agent-platform | Beginner | 0 | 0 | 110 | 60 | n/a |