Model Details
Longcat Flash Chat
Quick answer: Longcat Flash Chat is Meituan's July 2025 chat-optimised model scoring 88.4% HumanEval, 85.5% MMLU, and 82.7% MMLU-Pro. The chat variant of Longcat Flash (vs Thinking).
At a Glance
Where Longcat Flash Chat leads
- 88.4% HumanEval — excellent coding
- 82.7% MMLU-Pro — strong academic knowledge
- 85.5% MMLU — solid general knowledge
Where it lags
- Chat variant: no extended reasoning (vs Thinking variant)
- Proprietary: Meituan API
Best for: Meituan API chat deployments; high-HumanEval coding with MMLU-Pro knowledge.
What Longcat Flash Chat Is
Longcat Flash Chat is the conversation-optimised variant of Meituan's Longcat Flash, compared to the Longcat Flash Thinking variant which uses chain-of-thought reasoning. Chat is faster and better for low-latency applications.
Specifications
| Field | Value |
|---|---|
| Organization | Meituan |
| License | Proprietary |
| Release date | July 2025 |
| Modality | Text only |
Pricing
Available via Meituan AI API.
Public Benchmark Scores
| Benchmark | Score | Source | Date |
|---|---|---|---|
| HumanEval | 88.4% | Benchgen evaluation | 2025-07 |
| MMLU | 85.5% | Benchgen evaluation | 2025-07 |
| MMLU-Pro | 82.7% | Benchgen evaluation | 2025-07 |
Longcat Flash Chat vs Alternatives
| Model | HumanEval | MMLU-Pro | Type | License |
|---|---|---|---|---|
| Longcat Flash Chat | 88.4% | 82.7% | Chat | Proprietary |
| Longcat Flash Thinking | — | 82.6% | Thinking | Proprietary |
Chat vs Thinking: near-equal MMLU-Pro (82.7% vs 82.6%). Chat adds HumanEval (88.4%) but Thinking adds BrowseComp (56.6%) and LiveCodeBench (79.4%).
Related Models & Benchmarks
- Compare: Longcat Flash Thinking
- Benchmarks: HumanEval · MMLU-Pro
Frequently Asked Questions
What is Longcat Flash Chat?
Meituan's July 2025 chat variant scoring 88.4% HumanEval, 82.7% MMLU-Pro. Proprietary.Specs from Meituan's Longcat Flash Chat release (July 2025) and Benchgen evaluations. Last updated 2026-07-24.