MEITUAN LONGCAT1M CONTEXTAGENTIC CODINGCNY SOURCE
LongCat-2.0 API Pricing
LongCat-2.0 is Meituan's agentic coding model and the first Meituan entry in this catalogue. The vendor list price is CNY 5/M input, CNY 0.10/M cached input and CNY 20/M output, stored here as $0.70, $0.014 and $2.82 at the project CNY/USD rate of 7.10. A launch promotion currently cuts every rate by 60%, but it is explicitly temporary - the figures on this page are list price.
Input - per 1M tokens
$0.70/M
CNY 5 per 1M at 7.10 list price
Output - per 1M tokens
$2.82/M
CNY 20 per 1M at 7.10 list price
Cached input - per 1M tokens
$0.01/M
CNY 0.10 - unusually deep -98%
Effective - agentic blend
$0.35/M
92/8 split - 82% cache
§ 01 / TERMINAL
Run the numbers.
Live calculator pre-loaded with LongCat-2.0 list rates. Tweak the workload split, then share the URL to share the calculation.
$ /mo
Workload split
Prompt cache hit rate
Tokens you can process
—
Words equivalent (English)
—
Effective rate
—
§ 02 / SCENARIOS
Real-world presets.
CODING AGENT
Codebase-scale migration
$0.045/task
LONG DOC
Reading 100-page contracts
$0.047/doc
RAG SUPPORT
Ticket triage
$0.00392/ticket
ASSISTANT
Research planning turn
$0.00812/turn
§ 03 / TOKENIZER
Paste text. See tokens. See cost.
Estimate · meituan-tokenizer-estimate · ≈3.85 chars/token Auto-counts as you type
This is a chars-per-token approximation, not a real tokenizer. Actual tokens vary by language, code density, and tool-call overhead — counts are typically ±10–20% off for English prose, more for code or non-Latin scripts. For exact billing, use the vendor's official tokenizer.
Characters 595
Words 95
Tokens (estimated) 155 tokens
Cost as input · uncached $0.00011 USD
Cost as output · uncached $0.00044 USD
Cost as cached input $0.000002 USD
| Model | Input /M | Output /M | Effective blended | Context | Best for |
|---|---|---|---|---|---|
| LongCat-2.0 Current | $0.70 cache $0.01 | $2.82 | $0.35 agentic 92/8 | 1M | Agentic coding on Claude Code-compatible endpoints |
| Gemini 3.5 Flash-Lite | $0.30 cache $0.03 | $2.50 | $0.27 budget 1M peer | 1M | High-volume 1M-context work on a budget |
| Grok Build 0.1 | $1.00 cache $0.20 | $2.00 | $0.48 coding specialist | 256K | Coding agents with cheap output |
| Kimi K2.7 Code | $0.95 cache $0.19 | $4.00 | $0.62 Chinese coding peer | 262K | Repo-scale Chinese coding workloads |
| Muse Spark 1.1 | $1.25 cache $0.15 | $4.25 | $0.66 agentic 1M peer | 1M | Agentic tool and computer use |
| Kimi K3 | $3.00 cache $0.30 | $15.00 | $1.92 frontier agentic peer | 1M | Long-horizon agentic coding |
Frequently asked.
Practical LongCat-2.0 pricing questions, with the CNY list price separated from the temporary launch promotion.
Q · 01 What does LongCat-2.0 cost? +
Meituan's list price is
CNY 5 per 1M input tokens (cache miss), CNY 0.10 per 1M cached input tokens and CNY 20 per 1M output tokens. At the project rate of 7.10 CNY/USD that is $0.704225, $0.014085 and $2.816901.Q · 02 There is a cheaper price on the vendor page - why not use it? +
The pricing table prints a second column, 折扣价（限时）:
CNY 2 / CNY 0.04 / CNY 8, roughly 60% off. Meituan states it is a platform-launch limited-time offer and that actual charges are determined at settlement. We store list price so the row does not jump when the promotion lapses - but if you are buying today, the promotional rate is what you will be billed, and it puts the effective blend near $0.14/M instead of $0.35/M.Q · 03 How good is the cache discount? +
Unusually good. A cache hit costs
CNY 0.10 against CNY 5 on a miss - a 98% discount, deeper than the 88-90% typical elsewhere. No separate cache-write charge is published, so a miss bills at the normal input rate. Prompt reuse is worth more on this model than on most.Q · 04 What context window and output limit does it support? +
The docs list a
1M context window and a 128K maximum output length. The exact input token limit is not published as a number, so we do not record one.Q · 05 Can I use it with Claude Code or other agent harnesses? +
Yes. The platform serves both an OpenAI-compatible endpoint (
api.longcat.chat/openai) and an Anthropic-compatible endpoint (api.longcat.chat/anthropic). Meituan's change log names Claude Code, Hermes, OpenClaw, OpenCode and Kilo Code as adapted environments.Q · 06 Is pay-as-you-go the only billing mode? +
No. Meituan also sells a prepaid Token resource pack valid for 30 calendar days. That is a bundle, not a per-token rate, so it is not recorded here - every figure on this page is the pay-as-you-go rate.
Q · 07 How does it compare to other agentic models? +
At an effective
$0.35/M on list price it undercuts Kimi K2.7 Code by 76% and Muse Spark 1.1 by 87%. Only Gemini 3.5 Flash-Lite comes in cheaper among freshly verified 1M-context peers, at 23% less.Q · 08 What happened to the older LongCat models? +
Meituan retired six of them on
May 29, 2026 - LongCat-Flash-Chat, Flash-Thinking, Flash-Thinking-2601, Flash-Lite, Flash-Omni-2603 and Flash-Chat-2602-Exp - to concentrate resources on the 2.0 line. LongCat-2.0 is the only model the pay-as-you-go API serves today.