Qwen3-Coder-Next Technical Report

TL;DR

Qwen3-Coder-Next is an 80-billion-parameter coding language model that activates only 3 billion parameters during inference, demonstrating strong performance through innovative training methods and efficient inference.

Contribution

This work introduces a highly efficient coding model with a small active parameter count, trained via agentic methods and large-scale synthesis, pushing the limits of small-footprint models.

Findings

Achieves competitive performance on coding benchmarks.

Utilizes environment feedback for training.

Supports open research with released models.

Abstract

We present Qwen3-Coder-Next, an open-weight language model specialized for coding agents. Qwen3-Coder-Next is an 80-billion-parameter model that activates only 3 billion parameters during inference, enabling strong coding capability with efficient inference. In this work, we explore how far strong training recipes can push the capability limits of models with small parameter footprints. To achieve this, we perform agentic training through large-scale synthesis of verifiable coding tasks paired with executable environments, allowing learning directly from environment feedback via mid-training and reinforcement learning. Across agent-centric benchmarks including SWE-Bench and Terminal-Bench, Qwen3-Coder-Next achieves competitive performance relative to its active parameter count. We release both base and instruction-tuned open-weight versions to support research and real-world coding agent development.