← Niche Catalog

code_refactor

Modality: llm_chat · full deep dive — every ranked model, test result, and artifact.

13
Models
0
Benchmark Results
0
Media Artifacts
0
Resolutions

Ranked Models

Ranked by confidence-adjusted score (single/zero-sample, non-curated scores floored; curated empirical scores trusted as-is).

#ModelProviderAdj. ScoreRawEvidence
1claude-sonnet-4-6anthropic0.9200.920curated
2claude-opus-4-6anthropic0.8400.840curated
3deepseek-reasonerdeepseek0.7000.700curated
4gpt-5.5::think=fastcodex0.5250.963n=6
5gpt-5.5::think=deepcodex0.5250.963n=6
6gpt-5.3-codex-spark::think=deepcodex0.3820.700n=6
7gpt-5.4::think=fastcodex0.3340.612n=6
8gpt-5.4::think=deepcodex0.3340.612n=6
9codex-auto-reviewcodex0.2860.525n=6
10gpt-5.3-codex-spark::think=fastcodex0.2860.525n=6
11gpt-5.4codex0.0000.000n=50
12gpt-5.5codex0.0000.000n=50
13gpt-5.3-codex-sparkcodex0.0000.000n=50

Test Results

No benchmark outputs recorded for this niche yet.