Modality: llm_chat · full deep dive — every ranked model, test result, and artifact.
Ranked by confidence-adjusted score (single/zero-sample, non-curated scores floored; curated empirical scores trusted as-is).
| # | Model | Provider | Adj. Score | Raw | Evidence |
|---|---|---|---|---|---|
| 1 | claude-sonnet-4-6 | claude_desktop | 0.773 | 0.850 | n=50 |
| 2 | claude-sonnet-4-6::think=fast | claude_code | 0.545 | 1.000 | n=6 |
| 3 | claude-sonnet-4-6::think=deep | claude_code | 0.545 | 1.000 | n=6 |
| 4 | gpt-5.5::think=fast | codex | 0.525 | 0.963 | n=6 |
| 5 | gpt-5.5::think=deep | codex | 0.525 | 0.963 | n=6 |
| 6 | gpt-5.3-codex-spark::think=deep | codex | 0.382 | 0.700 | n=6 |
| 7 | gpt-5.4::think=fast | codex | 0.334 | 0.612 | n=6 |
| 8 | gpt-5.4::think=deep | codex | 0.334 | 0.612 | n=6 |
| 9 | gpt-5.3-codex-spark::think=fast | codex | 0.286 | 0.525 | n=6 |
| 10 | gemini-3.1-pro | gemini | 0.120 | 0.800 | provisional (n≤1) |
| 11 | claude-sonnet-4-6 | claude_code | 0.000 | 0.000 | n=50 |
| 12 | claude-sonnet-4-5-20250929 | claude_code | 0.000 | 0.000 | n=50 |
| 13 | claude-sonnet-4-20250514 | claude_code | 0.000 | 0.000 | n=50 |
| 14 | gpt-5.4 | codex | 0.000 | 0.000 | n=50 |
| 15 | gpt-5.5 | codex | 0.000 | 0.000 | n=50 |
| 16 | gpt-5.3-codex-spark | codex | 0.000 | 0.000 | n=50 |
No benchmark outputs recorded for this niche yet.