# Kwaipilot Releases KAT-Coder V2.5 Dev, an Agentic MoE Coder

> Kuaishou's coding team ships an open mixture-of-experts model built on the Qwen3.5 MoE architecture and tuned for agentic development work.

Published by The Open Weights on Jul 23, 2026. Canonical: https://theopenweights.com/news/kat-coder-v2-5-dev-tv94

## Key facts

- Company: Kwaipilot
- Model: KAT-Coder V2.5 Dev
- Version: 2.5
- Category: Code
- Modalities: Code, Text / LLM
- License: Other (Open weights, commercial use allowed)
- Context window: 262K tokens
- Architecture: Qwen3_5MoeForConditionalGeneration
- Precision: BF16
- Weights size: 69.3 GB
- Significance: notable
- Published: 2026-07-23
- Last verified: 2026-08-13
- Hugging Face: https://huggingface.co/Kwaipilot/KAT-Coder-V2.5-Dev
- Canonical URL: https://theopenweights.com/news/kat-coder-v2-5-dev-tv94

Kwaipilot, the developer-tools group inside Kuaishou, has published [KAT-Coder V2.5 Dev](https://huggingface.co/Kwaipilot/KAT-Coder-V2.5-Dev) on Hugging Face. The model is aimed squarely at agentic coding — the kind of multi-step, tool-using workflows where an assistant plans, edits, and iterates on code rather than simply completing a single prompt.

Under the hood, KAT-Coder V2.5 Dev is a mixture-of-experts (MoE) model built on the Qwen3.5 MoE architecture. MoE designs route each token through a subset of specialized expert networks, which lets a model carry a large total parameter count while keeping the compute cost of any single forward pass lower than a comparably sized dense model. That trade-off is increasingly popular for code models, where breadth of knowledge matters but latency and cost still bite.

## Why it matters

The coding-model space has become one of the most competitive corners of open-weights AI, and Kuaishou's entry adds another well-resourced backer to the field.

- It leans on the widely adopted Qwen MoE lineage, which lowers the barrier for tooling and integration.
- Its focus on agentic workflows aligns with where developer tools are heading, from single-shot completion toward autonomous task execution.
- As an openly published model, it gives teams an alternative to closed coding assistants for self-hosted or customized deployments.

Some key details remain unstated in the release, including exact parameter counts, active-expert configuration, and context length. The model is distributed under a custom license, so teams evaluating it for production should review the terms on the [model page](https://huggingface.co/Kwaipilot/KAT-Coder-V2.5-Dev) before building on top of it.

## Get the model

- [Hugging Face](https://huggingface.co/Kwaipilot/KAT-Coder-V2.5-Dev)

## Sources

- [Kwaipilot/KAT-Coder-V2.5-Dev](https://huggingface.co/Kwaipilot/KAT-Coder-V2.5-Dev) — Hugging Face, Jul 23, 2026

---
Source: The Open Weights (https://theopenweights.com/). Aggregated and written by Claude, curated by humans. Cite as: "Kwaipilot Releases KAT-Coder V2.5 Dev, an Agentic MoE Coder", The Open Weights, Jul 23, 2026, https://theopenweights.com/news/kat-coder-v2-5-dev-tv94