Kwaipilot: KAT-Coder-Pro V2
kwaipilot/kat-coder-pro-v2
72nd smartest of 182ToolsJSON
Intelligence
33.7
72nd of 182
Design Elo
—
Speed
—
tokens/sec
Latency
—
first token
Input price
$0.300
137th cheapest
Context
262K
80K max out
How it compares
Smarter than60%
of all ranked models
Cheaper than59%
of all ranked models
Overview
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,...
Benchmarks
independent · Artificial Analysis & Design ArenaArtificial Analysis81th percentile
Intelligence Index
33.7
Coding Index
59.5
GPQA Diamond
86%
Humanity's Last Exam
16%
SciCode
38%
Tau²-Bench (agentic)
89%
Providers & pricing (2)
| Provider | In $/M | Out $/M | Context | Uptime |
|---|---|---|---|---|
| StreamLake | $0.300 | $1.20 | 256K | — |
| AtlasCloudfp8 | $0.300 | $1.20 | 262K | — |
Specifications
Context window262K
Max output80K
Knowledge cutoff—
Input modalitiestext
Output modalitiestext
Prompt caching—
Cache read price$0.060/M
ModeratedNo
KAT-Coder-Pro V2 FAQ
How much does KAT-Coder-Pro V2 cost?
KAT-Coder-Pro V2 costs $0.300 per million input tokens and $1.20 per million output tokens via OpenRouter, making it 137th cheapest of 332 paid models.
How smart is KAT-Coder-Pro V2?
KAT-Coder-Pro V2 scores 33.7 on the Artificial Analysis Intelligence Index, ranking 72nd of 182 benchmarked models, with a GPQA Diamond score of 86%.
What is KAT-Coder-Pro V2's context window?
KAT-Coder-Pro V2 supports a 262K-token context window and can output up to 80K tokens. It accepts text input.
Similar models
All KAT-Coder-Pro V2 alternatives →Compare head-to-head
Kwaipilot: KAT-Coder-Pro V2 vs Qwen: Qwen3.5 397B A17BKwaipilot: KAT-Coder-Pro V2 vs MiniMax: MiniMax M2.5Kwaipilot: KAT-Coder-Pro V2 vs Z.ai: GLM 4.7Kwaipilot: KAT-Coder-Pro V2 vs Qwen: Qwen3.5-27BKwaipilot: KAT-Coder-Pro V2 vs Tencent: Hy3 previewKwaipilot: KAT-Coder-Pro V2 vs Meituan: LongCat 2.0