# LongCat-2.5 Benchmark Scores & Performance

> LongCat-2.5 by Meituan has 26 source-displayable benchmark rows but no public overall score. It remains unranked.

Canonical page: https://benchlm.ai/models/longcat-2-5

Last updated: November 9, 2026

## Model Details

| Property | Value |
|----------|-------|
| Creator | Meituan |
| Source Type | Open Weight |
| Reasoning Type | Hybrid |
| Context Window | 1M |
| Official model card | [Meituan LongCat-2.5 model card](https://huggingface.co/meituan-longcat/LongCat-2.5) |
| Overall Score | Not computed (unranked) |
| Overall Rank | Unranked |

## Family & Coverage

- Family: LongCat-2.5
- Variant: base
- Benchmarks covered: 26 of 788
- Related earlier model: [LongCat-2.0](/models/longcat-2-0)
- Coverage note: Published evidence is shown below; no public overall score is assigned.

## Agentic Benchmarks

| Benchmark | Score | Source |
|-----------|-------|--------|
| [Terminal-Bench 2.1](/benchmarks/terminalbench21) | 88.4% | [Meituan: LongCat-2.5 model card](https://huggingface.co/meituan-longcat/LongCat-2.5) |
| [Terminal-Bench 4.0](/benchmarks/terminal-bench) | 38.50% | [Meituan: LongCat-2.5 model card](https://huggingface.co/meituan-longcat/LongCat-2.5) |
| [BrowseComp](/benchmarks/browsecomp) | 91.9% | [Meituan: LongCat-2.5 model card](https://huggingface.co/meituan-longcat/LongCat-2.5) |
| [AutomationBench](/benchmarks/automationbench) | 51.8% | [Meituan: LongCat-2.5 model card](https://huggingface.co/meituan-longcat/LongCat-2.5) |
| [HLE w/ tools](/benchmarks/hlewithtools) | 63.1% | [Meituan: LongCat-2.5 model card](https://huggingface.co/meituan-longcat/LongCat-2.5) |
| [Agents' Last Exam](/benchmarks/agentslastexam) | 38.5% | [Meituan: LongCat-2.5 model card](https://huggingface.co/meituan-longcat/LongCat-2.5) |
| [MCP Atlas](/benchmarks/mcpatlas) | 83.1% | [Meituan: LongCat-2.5 model card](https://huggingface.co/meituan-longcat/LongCat-2.5) |
| [JobBench](/benchmarks/jobbench) | 65.4% | [Meituan: LongCat-2.5 model card](https://huggingface.co/meituan-longcat/LongCat-2.5) |
| [skillsBench](/benchmarks/skillsbench) | 63.3% | [Meituan: LongCat-2.5 model card](https://huggingface.co/meituan-longcat/LongCat-2.5) |
| [GDPval rubrics](/benchmarks/gdpvalrubrics) | 84.8% | [Meituan: LongCat-2.5 model card](https://huggingface.co/meituan-longcat/LongCat-2.5) |
| [SpreadsheetBench 2](/benchmarks/spreadsheetbench2) | 34.9% | [Meituan: LongCat-2.5 model card](https://huggingface.co/meituan-longcat/LongCat-2.5) |
| [Claw-Eval](/benchmarks/claw-eval) | 78.3% | [Meituan: LongCat-2.5 model card](https://huggingface.co/meituan-longcat/LongCat-2.5) |

## Coding Benchmarks

| Benchmark | Score | Source |
|-----------|-------|--------|
| [Terminal-Bench 2.1](/benchmarks/terminalbench21) | 88.4% | [Meituan: LongCat-2.5 model card](https://huggingface.co/meituan-longcat/LongCat-2.5) |
| [ProgramBench](/benchmarks/programbench) | 25.5% | [Meituan: LongCat-2.5 model card](https://huggingface.co/meituan-longcat/LongCat-2.5) |
| [DeepSWE](/benchmarks/deepswe) | 64.6% | [Meituan: LongCat-2.5 model card](https://huggingface.co/meituan-longcat/LongCat-2.5) |
| [PaperBench](/benchmarks/paperbench) | 82.1% | [Meituan: LongCat-2.5 model card](https://huggingface.co/meituan-longcat/LongCat-2.5) |
| [SWE-Atlas Codebase QnA](/benchmarks/swe-atlas-codebase-qna) | 58.3% | [Meituan: LongCat-2.5 model card](https://huggingface.co/meituan-longcat/LongCat-2.5) |
| [NL2Repo](/benchmarks/nl2repo) | 60.5% | [Meituan: LongCat-2.5 model card](https://huggingface.co/meituan-longcat/LongCat-2.5) |

## Multimodal & Grounded Benchmarks

| Benchmark | Score | Source |
|-----------|-------|--------|
| [Vision2Web](/benchmarks/vision2web) | 78.6% | [Meituan: LongCat-2.5 model card](https://huggingface.co/meituan-longcat/LongCat-2.5) |
| [CharXiv](/benchmarks/charxiv) | 90.2% | [Meituan: LongCat-2.5 model card](https://huggingface.co/meituan-longcat/LongCat-2.5) |
| [MathVision w/ Python](/benchmarks/mathvisionpython) | 96.5% | [Meituan: LongCat-2.5 model card](https://huggingface.co/meituan-longcat/LongCat-2.5) |
| [Video-MME](/benchmarks/videomme) | 85.2% | [Meituan: LongCat-2.5 model card](https://huggingface.co/meituan-longcat/LongCat-2.5) |

## Knowledge Benchmarks

| Benchmark | Score | Source |
|-----------|-------|--------|
| [FrontierScience](/benchmarks/frontierscience) | 77% | [Meituan: LongCat-2.5 model card](https://huggingface.co/meituan-longcat/LongCat-2.5) |

## Other Meituan Models

- [LongCat-Flash-Lite-Sparse](/models/longcat-flash-lite-sparse) - Score: not computed
- [LongCat-2.0](/models/longcat-2-0) - Score: not computed