Kwaipilot: KAT-Coder-Pro V2

Text input Text output
Author's Description

KAT-Coder-Pro V2 is the latest high-performance model in KwaiKAT’s KAT-Coder series, designed for complex enterprise-grade software engineering and SaaS integration. It builds on the agentic coding strengths of earlier versions, with a focus on large-scale production environments, multi-system coordination, and seamless integration across modern software stacks, while also supporting web aesthetics generation to produce production-grade landing pages and presentation decks.

Key Specifications
Cost
$$$
Context
256K
Released
Mar 27, 2026
Speed
Ability
Reliability
Supported Parameters

This model supports the following parameters:

Temperature Frequency Penalty Stop Tool Choice Max Tokens Top P Tools
Features

This model supports the following features:

Tools
Performance Summary

Kwaipilot's KAT-Coder-Pro V2, released on March 27, 2026, demonstrates strong performance as a high-performance model for enterprise software engineering and SaaS integration. It consistently ranks among the fastest models, placing in the 81st percentile across eight benchmarks, and offers competitive pricing, ranking in the 58th percentile. Notably, the model exhibits exceptional reliability with a 98% success rate, indicating minimal technical failures and consistent evaluable responses. In terms of specific benchmarks, KAT-Coder-Pro V2 achieves perfect accuracy in both Hallucinations (100%) and General Knowledge (100%), positioning it as a highly accurate model at its price point and speed. It also performs very well in Email Classification (99% accuracy, 87th percentile) and Mathematics (93% accuracy, 74th percentile). While its Reasoning (84% accuracy, 66th percentile) and Instruction Following (65% accuracy, 66th percentile) capabilities are solid, its performance in Ethics (94% accuracy, 23rd percentile) and Coding (81% accuracy, 40th percentile) suggests areas for potential improvement compared to other models. Overall, its strengths lie in factual recall, classification, and mathematical problem-solving, making it a robust tool for many complex tasks, despite some variability in ethical and coding proficiency.

Model Pricing

Current Pricing

Feature Price (per 1M tokens)
Prompt $0.3
Completion $1.2
Input Cache Read $0.06

Price History

Available Endpoints
Provider Endpoint Name Context Length Pricing (Input) Pricing (Output)
StreamLake
StreamLake | kwaipilot/kat-coder-pro-v2-20260327 256K $0.3 / 1M tokens $1.2 / 1M tokens
AtlasCloud
AtlasCloud | kwaipilot/kat-coder-pro-v2-20260327 262K $0.3 / 1M tokens $1.2 / 1M tokens
Benchmark Results
Benchmark Category Reasoning Strategy Free Executions Accuracy Cost Duration
Other Models by kwaipilot