CLBench Life - Digital Footprints & Daily-Life Records: 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 browsing, transaction and location records; higher is better.

Metric: Task solving rate (%) on the 45 digital-footprint and daily-life record 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 2029. 10 models tracked. Updated 7 Oct 2026.

Top models

#ModelSettingScoreOverall rank
1GPT-5.4High19.3#135
2Gemini 3.1 Pro (Preview)High15.6#99
3Qwen 3.5 PlusHigh15.6#242
4Claude Opus 4.6High13.3#113
5Grok 4.20High13.3#301
6Hy3-previewHigh12.2#322
7Kimi K2.5High11.9#266
8Seed 2.0 ProHigh11.1#222
9DeepSeek V3.2Thinking7.4#419
10MiniMax-M2.5Thinking7.4#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-digital-footprints-daily-life-records · How It Works · Data refreshed daily, snapshot 2026-11-21.