Skip to content
llmwaves

LongCat 2.0

Meituan · released Jul 20, 2026

Open weightsReasoningTool use

LongCat 2.0 is a sparse mixture-of-experts language model from Meituan, with 48B active parameters out of 1.6T total. It is suited for coding, repository-level changes, long-horizon problem solving, and agentic...

Specification

Context window
1.05M
Max output
262K
Knowledge cutoff
Not stated
Parameters
1775.6B
Licence
mit
Serving providers
1
Moderated
No
Uptime
100.0%

Intelligence

33.9

79th percentile

Coding

45.3

Coding Index

Agentic

Agentic Index

Output speed

27 t/s

Median across providers

Latency

2.2s

Time to first token

Cost per task

$0.11

Estimated

Benchmarks

Where the score comes from

The Intelligence Index is a composite. These are the underlying evaluations this model was actually measured on.

Evaluation scores

Percentage correct · higher is better

  • GPQA Diamond
    78.0%
  • AA-LCR (long context)
    62.7%
  • SciCode
    35.4%
  • Humanity's Last Exam
    33.7%
  • τ²-bench (Telecom)
    12.8%

An evaluation missing from this list was not run for this model — it is not a zero.

View as table
EvaluationScore
GPQA Diamond78.0%
AA-LCR (long context)62.7%
SciCode35.4%
Humanity's Last Exam33.7%
τ²-bench (Telecom)12.8%

Against its peers

Intelligence Index · this model highlighted, nearest peers in grey

Peers are the models sitting closest on the Intelligence Index — the set you would realistically choose between.

View as table
ModelIntelligence
M2.534.5
GLM 4.734.5
Hy3 preview34.4
Qwen3.5 397B A17B34.3
LongCat 2.033.9
KAT-Coder-Pro V233.9
Kimi K2 Thinking33.5
o3 Pro33.3

Percentile among all indexed models

Intelligence79th
Coding58th

Pricing

What it costs to run

List prices per million tokens, plus what one representative task works out to.

List price

Input / 1M tokens
$0.3
Output / 1M tokens
$1.2
Cached input / 1M
$0.006
Blended 3:1
$0.525

One task, estimated

$0.11

Input tokens
50,000
Output tokens
80,000
Profile
Reasoning

Estimated from list pricing: 50K input tokens plus 80K output tokens for reasoning models (25K for non-reasoning).

Serving providers

Speed and latency figures are medians across these providers, so a widely-served model reports a blend rather than any single endpoint.