LongCat-Flash-Thinking-2601: benchmark results
January 2026 refresh of Meituan's open LongCat Flash Thinking reasoning MoE with large-scale agentic RL training. Provider: Meituan. Released 2026-01-14. Access: Open.
Aggregate Rating 1686 ± 16, rank #229 of 1656 rated models, from 55 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| VitaBench | 29.3 | Cross-Scenario Avg@4 (%) | 85 |
| LLM Stats Score | 35.16 | LLM Stats Score (conservative rating) | 73.5 |
| LLM Stats (BrowseComp-zh) | 69 | Score (%) | 71.4 |
| SuperCLUE General (March 2026) - Agentic Task Planning | 67.94 | Score | 65.2 |
| AGI-Eval Community - Instruction Following | 44.09 | Accuracy (%) | 65 |
| AGI-Eval Community - Algorithmic Reasoning | 43.33 | Accuracy (%) | 64.3 |
| ZeroEval GPQA Diamond | 80.5 | GPQA Diamond Score | 61.1 |
| AGI-Eval Community - Learning (English) | 91.13 | Accuracy (%) | 59.4 |
| ClawProBench | 57.48 | Final Score (imported) | 58.9 |
| AGI-Eval Community - Objective Accuracy (Chinese) | 62.61 | Accuracy (%) | 58 |
| AGI-Eval Community - Subject Knowledge | 84.41 | Accuracy (%) | 57.9 |
| AGI-Eval Community - Objective Accuracy | 67.51 | Accuracy (%) | 57.1 |
Interactive version: theaggregate.ai/model?slug=longcat-flash-thinking-2601 · How It Works · Data refreshed daily, snapshot 2026-10-20.