CLBench Life - Creation & Revision Histories: leaderboard

CL-bench Life (2026) tests whether language models can learn from long, messy real-life context to solve human-written tasks, a task counting as solved only when every expert rubric item is met. This board is the task solving rate in percent on tasks over document edit histories and revision suggestions; higher is better.

Metric: Task solving rate (%) on the 45 creation and revision history tasks of CL-bench Life; a task is solved only when the response satisfies every human-written binary rubric item, judged by GPT-5.1 at high reasoning effort from the response and rubrics alone; reasoning mode at high effort where the model supports it, mean of three runs; higher is better. Source: arxiv.org. Saturation forecast: Around August 2028. 10 models tracked. Updated 7 Oct 2026.

Top models

#ModelSettingScoreOverall rank
1Claude Opus 4.6High20#113
2Gemini 3.1 Pro (Preview)High18.5#99
3Hy3-previewHigh14.4#322
4Kimi K2.5High13.3#266
5GPT-5.4High12.6#135
6Seed 2.0 ProHigh11.9#222
7Qwen 3.5 PlusHigh11.9#242
8Grok 4.20High9.6#301
9DeepSeek V3.2Thinking6.7#419
10MiniMax-M2.5Thinking4.4#600

No result here: #1 Claude Fable 5.2, #3 Gemini 4 Neon, #4 GPT-6.2 Astra.

Interactive version: theaggregate.ai/benchmark?slug=clbench-life-creation-revision-histories · How It Works · Data refreshed daily, snapshot 2026-11-21.