LongCat-2.0vsLongCat-2.5

Both language models, compared at each maker’s own API list prices. Every figure links to its source.

Compare at your usage

Every model priced for the same request
Prompt per request
Cache hits

8,000 tokens in, 2,000 out

LongCat-2.0 on LongCat APILongCat-2.5 on LongCat API
Rate appliedBase rateBase rate
Inputper 1M tokens$0.30 $0.30
Cache readper 1M tokens$0.006 $0.006
Outputper 1M tokens$1.20 $1.20
Cost per request$0.0048$0.0048
Per 1,000 requests$4.80$4.80

Every model is priced for the same request at its maker’s own list prices (each with its source and date), so bold marks the cheapest. A model that can’t serve the request is named, not ranked; a condition its seller doesn’t price says which rate was used.

Cost per request as prompts grow

2K tokens out, list price on each maker’s API.

0$0.10$0.20$0.300250K500K750K1Mprompt tokens per request8KLongCat-2.0$0.0048 at 8KLongCat-2.5$0.0048 at 8K
  • LongCat-2.0 · LongCat API
  • LongCat-2.5 · LongCat API

Key differences

Max output
131,072 LongCat API: chat completions · as of Nov 2, 2026
256,000 快速开始 | LongCat API Docsin Chinese · as of Nov 3, 2026

These differences use verified vendor specifications or each maker’s official API on the same region, currency, meter and billing conditions. Different conditions remain side by side in the full table; they do not establish which model is cheaper. Declared features describe support, not measured quality or a recommendation. Missing evidence does not mean a feature is unsupported.

Full comparison

LongCat-2.0 Meituan LongCat-2.5 Meituan
Overview
ReleasedJun 30, 2026 Nov 2, 2026
Open weightsYes —
Specs
Context1M 1M
Max output131.07K 256K
Input modalitiesText Not stated
Output modalitiesText Not stated
ReasoningYes Not stated
Reasoning can be offYes Not stated
Tool useYes Not stated
Pricing · per 1M tokens
Input price$0.30 $0.30
Output price$1.20 $1.20
Cache read price$0.006 $0.006
Blended price (3 input : 1 output)$0.525 $0.525

Headline prices may come from different channels, regions or tiers. Check each source before choosing a model; missing facts stay blank.

Model details and sources