Kwaipilot: KAT-Coder-Pro V1
Analysis Summary
KAT-Coder-Pro V1 is Kwaipilot's first-generation coding-focused model, with an intelligence index of 36 and strong math performance (94.7 math index). Its GPQA and MMLU Pro scores are competitive, and the 256K context window supports longer document and codebase tasks. Tool use and function calling are both supported.
However, the coding index of 18.3 is low relative to its intelligence score, and the agentic index of 48.8 is moderate. The V2 successor is now benchmarked and outperforms V1 on agentic tasks, so V1 is no longer the recommended choice within this family. Pricing at $0.207 input and $0.828 output is relatively high for the capability tier.
V1 remains usable for math-heavy reasoning and structured content tasks, but teams evaluating Kwaipilot's offering should move directly to V2 for better agentic and coding performance at comparable cost.
Assessed June 6, 2026
Editorial notes
KAT-Coder-Pro V1 from Kwaipilot shows strong math and reasoning scores with a 256K context window, but its coding index is limited and it has been superseded by the benchmarked V2 release.
Rankings consider pricing, capabilities, benchmarks, and real-world applicability and are refreshed as new models launch. Feedback?
Performance Profile
How Kwaipilot: KAT-Coder-Pro V1 compares
Kwaipilot: KAT-Coder-Pro V1 ranks #77 of 378 AI models we track for overall intelligence, #155 of 315 for coding, #85 of 289 for agentic tasks. Its 256K-token context window is larger than 71% of the models we list. At $0.21 per million input tokens it is cheaper than 49% of comparable models.
About Kwaipilot: KAT-Coder-Pro V1
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
Performance Indices
Source: Artificial Analysis
Benchmark Scores
Intelligence
Technical
Content
Benchmark data from Artificial Analysis and Hugging Face
How does Kwaipilot: KAT-Coder-Pro V1 stack up?
Compare side-by-side with other professional models.
Model Information
| OpenRouter ID |
kwaipilot/kat-coder-pro
|
| Provider | kwaipilot |
| Release Date | November 10, 2025 |
| Context Length | 256,000 tokens |
| Max Completion | 128,000 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
External Resources
Frequently asked questions about Kwaipilot: KAT-Coder-Pro V1
How much does Kwaipilot: KAT-Coder-Pro V1 cost?
Kwaipilot: KAT-Coder-Pro V1 costs $0.21 per million input tokens and $0.83 per million output tokens.
What is the context window of Kwaipilot: KAT-Coder-Pro V1?
Kwaipilot: KAT-Coder-Pro V1 has a context window of 256,000 tokens (256K).
Is Kwaipilot: KAT-Coder-Pro V1 good for coding?
On our coding benchmark index, Kwaipilot: KAT-Coder-Pro V1 ranks #155 of 315 models, placing it in the broader range of the field for code generation and debugging.
What can Kwaipilot: KAT-Coder-Pro V1 do?
Kwaipilot: KAT-Coder-Pro V1 supports tool use and function calling.
Who created Kwaipilot: KAT-Coder-Pro V1?
Kwaipilot: KAT-Coder-Pro V1 is developed by Kwaipilot and was released on November 10, 2025.
