LiveBench Consecutive Events: leaderboard
Data analysis task in LiveBench, added in January 2026, that tests detecting specific events: the model returns a JSON list inside solution tags pairing user identifiers with each user's maximum number of consecutive growth months. It is scored by Jaccard overlap with the reference list, an entry counting only when identifier and count both match, with no LLM judge. The score runs from 0 to 100, higher is better.
Metric: Score. Source: livebench.ai. Saturation forecast: Around January 2027. 71 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Muse Wonder (Max) | 98.53 | #15 (Muse Wonder) |
| 2 | Claude Fable 5 (Max) | 91.36 | #16 (Claude Fable 5) |
| 3 | GPT-6.1 Sol (xHigh) | 91.29 | #7 (GPT-6.1 Sol) |
| 4 | Claude Sonnet 5.5 (xHigh) | 90.94 | #11 (Claude Sonnet 5.5) |
| 5 | GPT-6 Astra (Max) | 90.56 | #6 (GPT-6 Astra) |
| 6 | GPT-6.1 Sol (Max) | 90.56 | #7 (GPT-6.1 Sol) |
| 7 | Claude Fable 5.1 (Max) | 90.56 | #9 (Claude Fable 5.1) |
| 8 | Claude Opus 5.5 (Max) | 90.56 | #3 (Claude Opus 5.5) |
| 9 | Muse Spark 1.3 (xHigh) | 90.5 | #24 (Muse Spark 1.3) |
| 10 | Claude Opus 5.5 (xHigh) | 90.37 | #3 (Claude Opus 5.5) |
| 11 | DeepSeek V4 Pro (0813) | 90.23 | #53 |
| 12 | GPT-5.6 Sol (Max) | 90.21 | #19 (GPT-5.6 Sol) |
| 13 | DeepSeek-V4-Flash-Vision-Exp | 90.05 | #76 |
| 14 | Kimi K3 | 89.76 | #30 |
| 15 | GPT-6 Sol (Max) | 89.55 | #14 (GPT-6 Sol) |
No result here: #8 Gemini 4 Argon, #10 Muse Wonder Full.
Interactive version: theaggregate.ai/benchmark?slug=livebench-consecutive-events · How It Works · Data refreshed daily, snapshot 2026-10-20.