Private AI

Nano GPT logo
NanoGPT
Nano GPT logo
NanoGPT
Nano GPT logo
NanoGPT
Search
 
 
 
 
All conversations

Create

Chat
Studio

Explore

Models
Pricing
AI Apps
Blog

Library

Gallery
Workspace
Usage
API
Suggestions and bugs
Terms

|

Privacy

|

Refunds
Balance
Subscription
Teams
Referrals
Presets
Settings
FAQPress & mediaBrand assetsBenchmarksEarn crypto
  1. Models
  2. Text

LongCat 2.0 Thinking

Meituan's LongCat 2.0 Thinking is the reasoning-enabled variant for harder coding, tool use, multi-step reasoning, and long-context agent workflows.

  • Reasoning
  • Tool Calling

Added Jul 21, 2026

Model weights

Pricing

Auto routing · per 1M tokens
Input
$0.75
Output
$3.00
Cache read
$0.015
Compare provider prices

Specifications

Context window
1M
Max output
131.1K
Parameters
1.6T / 48B
Total / active
Avg output (7d)
1.3K tokens
Longer than 82% of models

Benchmarks

Sourced from Artificial Analysis.

Agentic work

AutomationBench-AA

Workflow automation with guardrail penalties

4.5%

Better than 24% of models compared

AutomationBench-AA Tasks Completed

Fully completed workflows without guardrail violations

0.5%

Better than 26% of models compared

AA-Briefcase v1.1

Agentic knowledge work (Elo)

740 Elo

Better than 28% of models compared

GDPval-AA v2.1

Economically valuable tasks (Elo)

877 Elo

Better than 36% of models compared

Document reasoning

GDP.pdf

Professional PDF reasoning: all-pass rate

6.4%

Better than 30% of models compared

AA-LCR v1.1

Long context reasoning with updated grading

65.0%

Better than 54% of models compared

Reasoning

HLE

Humanity's Last Exam

33.7%

Better than 79% of models compared

Coding

Terminal-Bench v4.0

Practical coding and terminal tasks

0.0%

Better than 15% of models compared

SciCode

Python programming for scientific computing

36.3%

Better than 15% of models compared

Legacy benchmarks

GPQA Diamond (legacy)

Graduate-level scientific reasoning

78.0%

Better than 63% of models compared

AA-LCR (unversioned / legacy)

Long context reasoning evaluation

65.0%

Better than 54% of models compared

Last updated Oct 30, 2026

Artificial Analysis

Providers

Loading provider options…

Related text models

Compare LongCat 2.0 Thinking with similar models from the same provider or model family.

LongCat 2.5 Preview

meituan/longcat-2.5-preview

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.

LongCat 2.0

longcat-2.0

Meituan's LongCat 2.0 is an open-weight agentic model for coding, tool use, and long-context workflow automation, with reasoning disabled for faster direct answers.

OrcaCyber Zero 1.5

orca/orcacyber-zero-1.5

OrcaRouter's frontier cybersecurity model, post-trained for authorized security research: vulnerability research and reproduction, exploit development, penetration testing, and security auditing. Supports a 1M-token context window, optional reasoning, tool calling, and structured output. Offensive work is permitted only within an authorized security engagement.

Particle-1

particle-ai/particle-1

Particle-1 is Particle AI’s coding and data-engineering model, post-trained on DeepSeek V4.1 Flash. It accepts text and images, supports tool calling, and offers a shared 1,048,576-token context window with up to 131,072 output tokens, including reasoning. Reasoning defaults to high and can be turned off. JSON formatting is supported; strict JSON-schema enforcement is not guaranteed.

DeepSeek V4.1 Flash 0910 Uncensored

deepseek/deepseek-v4.1-flash-0910-uncensored

Uncensored DeepSeek V4.1 Flash with text and image input, optional reasoning, and tool calling. The date label is retained for compatibility; the served checkpoint can vary with availability.

DeepSeek V4.1 Flash 0910 Uncensored Thinking

deepseek/deepseek-v4.1-flash-0910-uncensored:thinking

Uncensored DeepSeek V4.1 Flash with text and image input, optional reasoning, and tool calling. The date label is retained for compatibility; the served checkpoint can vary with availability.