Kwaipilot: KAT-Coder-Pro V1

Kwaipilot: KAT-Coder-Pro V1

kwaipilot · Released Nov 10, 2025 Specialist
Intelligence #73 / 525
65.3 Our Score
Speed #106 / 244
112.7 tokens / sec
Input #268 / 525
$0.207 per 1M tokens
Output #282 / 525
$0.828 per 1M tokens
Context #126 / 525
256,000 tokens

Analysis Summary

Kwaipilot: KAT-Coder-Pro V1 sits in the Specialist tier on our leaderboard, ranked #73 of 525 published models on overall intelligence. At $0.207 input and $0.828 output per 1M tokens, it is among the most expensive on the market. It offers a generous context window for extended reasoning and code review and supports tool use and function calling.

Editorial notes

Kwaipilot's KAT-Coder-Pro V1 shows strong instruction-following, long-context reasoning, and agentic scores at a competitive price, with tool use and function calling support. It is superseded by the V2 release for highlight purposes, but remains a capable model for content and workflow automation tasks.

Assessed April 23, 2026

Rankings consider pricing, capabilities, benchmarks, and real-world applicability and are refreshed as new models launch. Feedback?

Performance Profile

Intelligence6.9Technical4.7Value8Content7
Intelligence 6.9/10
Technical 4.7/10
Content 7/10
Value 8/10

KAT-Coder-Pro V1 is KwaiKAT's most advanced agentic coding model in the KAT-Coder series. Designed specifically for agentic coding tasks, it excels in real-world software engineering scenarios, achieving 73.4% solve rate on the SWE-Bench Verified benchmark. The model has been optimized for tool-use capability, multi-turn interaction, instruction following, generalization, and comprehensive capabilities through a multi-stage training process, including mid-training, supervised fine-tuning (SFT), reinforcement fine-tuning (RFT), and scalable agentic RL.

Capabilities

Tool Use Function Calling

Performance Indices

Source: Artificial Analysis

36 Intelligence Index
18.3 Coding Index
48.8 Agentic Index
94.7 Math Index

Benchmark Scores

Intelligence

GPQA Diamond 76.4% Graduate-level scientific reasoning
HLE 33.4% Humanity's Last Exam
MMLU Pro 81.3% Multi-task language understanding
AIME 2025 94.7% Competition mathematics (2025)
SciCode 36.6% Scientific computing

Technical

LiveCodeBench 74.7% Live coding evaluation
TerminalBench Hard 9.1% Agentic terminal tasks
τ²-Bench 88.6% Conversational agent benchmark

Content

IFBench 68.4% Instruction following
LCR 74% Long-context reasoning

Benchmark data from Artificial Analysis and Hugging Face

How does Kwaipilot: KAT-Coder-Pro V1 stack up?

Compare side-by-side with other specialist models.

Compare Models

Model Information

OpenRouter ID kwaipilot/kat-coder-pro
Providerkwaipilot
Release Date November 10, 2025
Context Length256,000 tokens
Max Completion128,000 tokens
Status Active

Pricing

Token Type Cost per 1M tokens Cost per 1K tokens
Input $0.21 $0.000207
Output $0.83 $0.000828

Leaderboard Categories