# J-WorkBench Report

- results: 2026-06-06
- runs: 7

## Leaderboard

| rank | model | kind | total | substitution | completed |
| ---: | --- | --- | ---: | ---: | ---: |
| 1 | qwen2.5:14b | local | 91 | 86 | 30/30 |
| 2 | gpt-oss:20b | local | 89 | 87 | 30/30 |
| 3 | qwen2.5-coder:14b | local | 83 | 68 | 30/30 |
| 4 | qwen2.5:7b | local | 82 | 66 | 29/30 |
| 5 | Codex (ChatGPT) | cloud | 79 | 100 | 30/30 |
| 6 | Claude (Claude Code / Max) | cloud | 73 | 100 | 29/30 |
| 7 | Gemini (Gemini CLI) | cloud | 72 | 100 | 30/30 |

## Audit Queue

- none
