CellTrack-Bench - Repair F1: leaderboard
CellTrack-Bench:frozen c1 hard-test subset,244 neighborhoods; same unannotated frame-pair coordinate interface,crop-and-zoom tool,prompt,parser and token budget. Dataset-macro scores. Pinned primary test configs use GPT5.5/GPT5.6 and Fable5/Opus4.8/Opus5/Sonnet5 Low with final reasoning None after tool withdrawal; Haiku4.5 uses dated20251001 endpoint and None. Open chat runner does not send configured reasoning_effort, so cached Medium is not an evaluated setting. Gemma4 is IT; Qwen3.6 result averages two same-checkpoint replicates. Nearest-neighbor/division resistance and graph-level correctness remain separate metrics. Score (0–1).
Metric: Score (0–1). Source: arxiv.org. Saturation forecast: Around January 2027. 22 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | GPT-5.6 Sol (Low) | 0.82 | #20 (GPT-5.6 Sol) |
| 2 | GPT-5.6 Terra (Low) | 0.79 | #39 (GPT-5.6 Terra) |
| 3 | Claude Fable 5 (Low) | 0.79 | #17 (Claude Fable 5) |
| 4 | GPT-5.5 (Low) | 0.74 | #30 (GPT-5.5) |
| 5 | Claude Opus 5 (Low) | 0.74 | #13 (Claude Opus 5) |
| 6 | Claude Opus 4.8 (Low) | 0.73 | #36 (Claude Opus 4.8) |
| 7 | GPT-5.6 Luna (Low) | 0.7 | #73 (GPT-5.6 Luna) |
| 8 | Kimi K3 | 0.66 | #32 |
| 9 | Qwen 3.5 397B A17B | 0.66 | #148 |
| 10 | MiniMax-M3 | 0.66 | #147 |
| 11 | Qwen 3.6 27B | 0.63 | #174 |
| 12 | Claude Sonnet 5 (Low) | 0.63 | #60 (Claude Sonnet 5) |
| 13 | Claude Haiku 4.5 (20251001) | 0.63 | #254 |
| 14 | Qwen 3.5 35B A3B | 0.62 | #259 |
| 15 | Gemma 4 31B (IT) | 0.61 | #179 |
No result here: #4 Claude Opus 5.5, #6 GPT-6 Astra, #7 GPT-6.1 Sol.
Interactive version: theaggregate.ai/benchmark?slug=celltrack-bench-repair-f1 · How It Works · Data refreshed daily, snapshot 2026-10-24.