NOUMENOpen app
NOUMEN

longcat · Balanced

LongCat 2.5 Preview

Added Sep 27, 2026

Compare models

Choose up to 3 models to compare.1 of 3 chosen

Loading models…

Use in the app

LongCat 2.5 Preview is Meituan's multimodal model for long-context analysis, coding, and agent workflows. It supports image input, tool calling, optional extended thinking, a 1M-token context window, and up to 128K output tokens.

  • Reasoning
  • Tools
  • Vision
Input
Text, Images
Output
Text
Discovery tier
Balanced

Pricing

USD per 1M tokens
Input
$0.75
Output
$2.95
Cache read
$0.015

Reference prices as published by the model provider. API example

Specifications

Context window
1M
Max output
131.1K

Side by side

LongCat 2.5 Preview compared

Specifications, prices, speed and benchmark scores next to each other. Benchmarks below show every compared model too.

Compare models

Choose up to 3 models to compare.1 of 3 chosen

Loading models…

LongCat 2.5 PreviewlongcatThis modelLongCat 2.0 ThinkinglongcatRemove
Providerlongcatlongcat
Discovery tierBalancedOpen weights
AddedSep 27, 2026Jul 21, 2026
Limits
Context window1M1M
Max output131.07K131.07K
Average output—1.1K tokens
Pricing, USD per 1M tokens
Input$0.75$0.75
Output$2.95Best$3.00
Cache read$0.015$0.015
Performance
Throughput—71.2 tok/s
Time to first token—6.2 s
Request success (24h)99.7%100.0%Best
Benchmark indexes
Intelligence Index—19.1
Coding Index—45.3
Capabilities
ReasoningYesYes
ToolsYesYes
VisionYesNo
Reasoning effort——

Public evaluations

Benchmarks

Independent evaluation results, with how each score ranks among the models in the Noumen catalog.

  • LongCat 2.5 Preview
  • LongCat 2.0 Thinking

Intelligence Index

  • LongCat 2.5 Preview—
  • LongCat 2.0 Thinking19.1
3.8Catalog models by score55.1

Indexes

  • Coding Index

    Composite of the coding evaluations

    • LongCat 2.5 Preview: —
    • LongCat 2.0 Thinking: 45.3

Agentic work

  • AutomationBench-AA

    Workflow automation, penalized for guardrail violations

    • LongCat 2.5 Preview: —
    • LongCat 2.0 Thinking: 4.5%

  • AutomationBench-AA tasks completed

    Workflows finished end to end without a guardrail violation

    • LongCat 2.5 Preview: —
    • LongCat 2.0 Thinking: 0.5%

  • AA-Briefcasev1.1

    Agentic knowledge work, rated head to head

    • LongCat 2.5 Preview: —
    • LongCat 2.0 Thinking: 739 Elo

  • GDPval-AAv2.1

    Economically valuable occupational tasks, rated head to head

    • LongCat 2.5 Preview: —
    • LongCat 2.0 Thinking: 869 Elo

Document reasoning

  • GDP.pdf

    Professional reasoning over PDF documents, all-pass rate

    • LongCat 2.5 Preview: —
    • LongCat 2.0 Thinking: 6.4%

  • AA-LCRv1.1

    Reasoning across long documents

    • LongCat 2.5 Preview: —
    • LongCat 2.0 Thinking: 65.0%

Reasoning and knowledge

  • Humanity’s Last Exam

    Expert-level questions across academic fields

    • LongCat 2.5 Preview: —
    • LongCat 2.0 Thinking: 33.7%

  • GPQA Diamond

    Graduate-level science questions

    • LongCat 2.5 Preview: —
    • LongCat 2.0 Thinking: 78.0%

Coding

  • Terminal-Benchv4.0

    Practical software engineering in a terminal

    • LongCat 2.5 Preview: —
    • LongCat 2.0 Thinking: 0.0%

  • SciCode

    Scientific computing in Python

    • LongCat 2.5 Preview: —
    • LongCat 2.0 Thinking: 36.3%

Legacy evaluations

  • AA-LCR (unversioned)

    Long-context reasoning, earlier grading

    • LongCat 2.5 Preview: —
    • LongCat 2.0 Thinking: 65.0%

Scores from Artificial Analysis, updated Oct 26, 2026. Artificial Analysis

Live routing data

Performance and reliability

Recent averages across requests routed for this model. Your own speed depends on prompt length, reasoning effort and load.

Prompt cache hits
79.0%
About $0.17 per 1M input tokens with caching
Request success (24h)
99.7%
Reliable

Cache hits are the share of input tokens recently served from a prompt cache across all requests. They are not a forecast of your own hit rate.

Hourly request success · last 72 hoursLowest hour 78.5%
  • Healthy (98% or more)
  • Degraded (90–98%)
  • Down (under 90%)
  • No requests

Developers

Use it through the API

Noumen’s API is OpenAI-compatible: any Chat Completions client works with a base URL and a key.

Read the quickstart
curl https://www.noumen.xyz/api/v1/chat/completions \
  -H "Authorization: Bearer $NOUMEN_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "meituan/longcat-2.5-preview",
    "messages": [{ "role": "user", "content": "Explain this stack trace." }]
  }'
import os
from openai import OpenAI

client = OpenAI(base_url="https://www.noumen.xyz/api/v1", api_key=os.environ["NOUMEN_API_KEY"])
reply = client.chat.completions.create(
    model="meituan/longcat-2.5-preview",
    messages=[{"role": "user", "content": "Explain this stack trace."}],
)
print(reply.choices[0].message.content)

longcat

Other longcat models, starting with the same family.