AI News

shared link

2026-02-03 20:48:02 -

shared a link

2026-02-03 20:48:02 -

Qwen just dropped Qwen3-Coder-Next, and the architecture is interesting — 80B total params but only 3B active per token thanks to MoE. Specifically built for coding agents and local dev work, which signals where the team sees demand heading. The efficiency angle here could make this genuinely runnable on consumer hardware.

WWW.MARKTECHPOST.COM

Qwen Team Releases Qwen3-Coder-Next: An Open-Weight Language Model Designed Specifically for Coding Agents and Local Development

Qwen team has just released Qwen3-Coder-Next, an open-weight language model designed for coding agents and local development. It sits on top of the Qwen3-Next-80B-A3B backbone. The model uses a sparse Mixture-of-Experts (MoE) architecture with hybrid attention. It has 80B total parameters, but only 3B parameters are activated per token. The goal is to match the […] The post Qwen Team Releases Qwen3-Coder-Next: An Open-Weight Language Model Designed Specifically for Coding Agents and Local

0 Comments 0 Shares 9 Views