# LongCat-Flash-Lite-Sparse Benchmark Scores & Performance

> LongCat-Flash-Lite-Sparse by Meituan has 17 source-displayable benchmark rows but no public overall score. It remains unranked.

Canonical page: https://benchlm.ai/models/longcat-flash-lite-sparse

Last updated: September 10, 2026

## Model Details

| Property | Value |
|----------|-------|
| Creator | Meituan |
| Source Type | Open Weight |
| Reasoning Type | Reasoning |
| Context Window | 1M |
| Official model card | [Meituan LongCat-Flash-Lite-Sparse model card](https://huggingface.co/meituan-longcat/LongCat-Flash-Lite-Sparse) |
| Overall Score | Coming soon |
| Overall Rank | Unranked |

## Family & Coverage

- Family: LongCat-Flash-Lite
- Variant: sparse
- Benchmarks covered: 17 of 434
- Coverage note: BenchLM currently has partial benchmark coverage for this model, so the overall score is conservative.

## Agentic Benchmarks

| Benchmark | Score |
|-----------|-------|
| [Terminal-Bench 2.0](/benchmarks/terminal-bench-2) | 33.7% |
| [VITA-Bench](/benchmarks/vitabench) | 21.7% |
| [MCP Atlas](/benchmarks/mcpatlas) | 45.6% |
| [BrowseComp](/benchmarks/browsecomp) | 48.6% |

## Coding Benchmarks

| Benchmark | Score |
|-----------|-------|
| [SWE-bench Verified](/benchmarks/swe-bench-verified) | 68.2% |
| [SWE-bench Pro](/benchmarks/swe-bench-pro) | 40.6% |
| [SWE Multilingual](/benchmarks/swe-bench-multilingual) | 59.3% |
| [Terminal-Bench 2.0](/benchmarks/terminal-bench-2) | 33.7% |

## Knowledge Benchmarks

| Benchmark | Score |
|-----------|-------|
| [MMLU](/benchmarks/mmlu) | 85.3% |
| [MMLU-Pro](/benchmarks/mmlu-pro) | 79.2% |
| [CMMLU](/benchmarks/cmmlu) | 84.3% |
| [C-Eval](/benchmarks/ceval) | 85.8% |
| [GPQA-D](/benchmarks/gpqa-diamond) | 69.5% |

## Mathematics Benchmarks

| Benchmark | Score |
|-----------|-------|
| [MATH-500](/benchmarks/math-500) | 95.8% |
| [AIME26](/benchmarks/aime2026) | 65.7% |
| [HMMT Feb 2026](/benchmarks/hmmtfeb2026) | 40.5% |
| [IMOAnswerBench](/benchmarks/imoanswerbench) | 49.4% |

## Other Meituan Models

- [LongCat-2.0](/models/longcat-2-0) - Score: not computed