LongCat-Flash-Thinking: API Pricing, Context Window & Benchmarks
LongCat-Flash-Thinking is a language model from Meituan, released in September 2025.
LongCat-Flash-Thinking is Meituan's reasoning model built on the LongCat-Flash foundation with 560B total parameters (MoE, ~27B activated). It introduces a training pipeline specifically tuned for advanced reasoning, featuring Re-thinking
LongCat-Flash-Thinking benchmarks
Rankings
Quality Tracker
LongCat-Flash-Thinking Performance Across Datasets
Scores sourced from the model's scorecard, paper, or official blog posts
llm-stats.com - Sat Jul 18 2026LongCat-Flash-Thinking pricing
Providers
LongCat-Flash-Thinking starts at $0.300 per million input tokens and $1.20 per million output tokens via Meituan.
| Provider | Input $/M | Output $/M | Workload 1M + 100K | Context in / out | TTFT p50 / p95 s | Output avg / p5 c/s | Success 7d | Modalities in / out |
|---|---|---|---|---|---|---|---|---|
 Meituan | $0.300 | $1.20 | $0.420 | 128.0K/128.0K | —/3.00 | 100/— | — | / |
Workload cost uses 1M input tokens plus 100K output tokens at published list prices without assuming a cache hit. Cached input is the discounted price for prompt tokens served from a provider cache. TTFT is time to first token. Output is characters per second; p5 is the sustained floor exceeded by 95% of observed requests. Success is calculated from completed versus failed requests over the trailing seven days.
LongCat-Flash-Thinking model size
LongCat-Flash-Thinking has 560 billion parameters. See how it compares to other models in the same parameter range.
LongCat-Flash-Thinking context window
Input and output token limits for LongCat-Flash-Thinking, plus how it ranks on long-context understanding.
LongCat-Flash-Thinking API
Available from the model provider
LongCat-Flash-Thinking has an official provider API. It is not currently routed through the LLM Stats gateway.
Read the official API documentationLongCat-Flash-Thinking latency
LongCat-Flash-Thinking time to first token, sustained output throughput, and failed-request rate from live API traffic over the trailing 7 days.
LongCat-Flash-Thinking examples
Recent arena outputs from LongCat-Flash-Thinking, picked from the highest-ranked matchups.
LongCat-Flash-Thinking license
LongCat-Flash-Thinking is released under the MIT license, which permits commercial use, has 560.0B parameters.
- License
- MIT
- Commercial use allowed
- Parameters
- 560.0B
MIT License - allows commercial use
LongCat-Flash-Thinking resources
Official sources for LongCat-Flash-Thinking: api documentation, official playground, paper or system card, source repository, model weights.
LongCat-Flash-Thinking vs other models
The most-compared alternatives to LongCat-Flash-Thinking are Grok-3 Mini, Kimi K2 0905, Nova 2 Pro. Open any pair side-by-side for benchmarks, pricing, context, and latency.
Models like LongCat-Flash-Thinking
Models ranked just above and below LongCat-Flash-Thinking by LLM Stats score.
FAQ
Common questions about LongCat-Flash-Thinking.