| Metric | Value |
|---|---|
| GitHub Copilot Workspace wins | 4 |
| e2b wins | 4 |
| Abstains (no tool) | 348 |
| Other tool chosen | 868 |
| Decisive cases | 8 |
| GitHub Copilot Workspace win rate (unweighted) | 50.0% |
| 95% CI | 21.5% - 78.5% |
| GitHub Copilot Workspace win rate (weighted) | 50.0% |
Verified critics can leave comments here.
Verified critics can leave comments here.
| Model | Tier | GitHub Copilot Workspace | e2b | None | Other | A rate |
|---|---|---|---|---|---|---|
| Gemini 2.5 Pro | Frontier | 1 | 2 | 9 | 54 | 33% |
| Qwen3 Coder Next | Mid | 0 | 2 | 8 | 55 | 0% |
| GPT 5.4 Mini | Mid | 1 | 0 | 4 | 61 | 100% |
| MiniMax M2.7 | Frontier | 1 | 0 | 11 | 52 | 100% |
| Mistral Small 4 | Mid | 1 | 0 | 28 | 28 | 100% |
| Claude Haiku 4.5 | Small | 0 | 0 | 32 | 30 | n/a |
| Claude Opus 4.6 | Frontier | 0 | 0 | 8 | 58 | n/a |
| Claude Sonnet 4.6 | Frontier | 0 | 0 | 28 | 38 | n/a |
| DeepSeek R1 0528 | Frontier | 0 | 0 | 26 | 40 | n/a |
| DeepSeek V3.2 | Mid | 0 | 0 | 25 | 41 | n/a |
| Devstral 2 2512 | Mid | 0 | 0 | 25 | 38 | n/a |
| Gemini 2.5 Flash | Small | 0 | 0 | 31 | 35 | n/a |
| GLM 5 Turbo | Frontier | 0 | 0 | 0 | 66 | n/a |
| GPT 5.3 Codex | Frontier | 0 | 0 | 13 | 53 | n/a |
| GPT 5.4 | Frontier | 0 | 0 | 13 | 50 | n/a |
| Kimi K2.5 | Frontier | 0 | 0 | 24 | 35 | n/a |
| Llama 4 Maverick | Frontier | 0 | 0 | 27 | 39 | n/a |
| Llama 4 Scout | Small | 0 | 0 | 21 | 44 | n/a |
| MiMo V2 Pro | Frontier | 0 | 0 | 15 | 51 | n/a |
| Prompt | Tier | GitHub Copilot Workspace | e2b | None | Other | A rate |
|---|---|---|---|---|---|---|
| ai-revenue-ops-copilot | Advanced | 2 | 2 | 2 | 188 | 50% |
| ai-support-agent-platform | Advanced | 1 | 2 | 0 | 203 | 33% |
| ai-revenue-ops-copilot | Beginner | 1 | 0 | 112 | 94 | 100% |
| ai-revenue-ops-copilot | Intermediate | 0 | 0 | 1 | 200 | n/a |
| ai-support-agent-platform | Beginner | 0 | 0 | 136 | 72 | n/a |
| ai-support-agent-platform | Intermediate | 0 | 0 | 97 | 111 | n/a |