← Harnesses
structured_feedback
legacysupersededEarly structured test-feedback experiment.
why it exists: Precursor to evidence-carrying retries.
why retired: Folded into v1.4. → v1.4
Models
4
have run it
Solved
23
gauntlet benches
Models on this harness (gauntlet solved / attempted)
| Model | Solved | By domain |
|---|---|---|
| gemma-4-26b-a4b-it | 6/6 | Coding 6/6 |
| gpt-oss-20b | 6/6 | Coding 6/6 |
| qwen3.6-35b-a3b | 6/6 | Coding 6/6 |
| qwen3.5-4b | 5/6 | Coding 5/6 |