Back to directory

Qwen3.8

The official repository for Alibaba's Qwen3.8 large language model, which brings a top-tier Qwen-Max-class model to open release, with stronger coding, professional-work, research and long-horizon agentic performance.

Qwen3.8

Qwen3.8 is the latest release in the Qwen3.5 open-model series from Alibaba's Qwen team, following Qwen3.5 and Qwen3.6, and is the first time a Qwen-Max-class model has been made available as an open release. Built on the architectural foundation of Qwen3.5, it improves coding, professional work, research and long-horizon agentic tasks, and is designed to carry complex, multi-step tasks through to completion more reliably rather than only answering harder individual questions.

Its agent-execution capability adds stronger autonomous planning and better handling of environment feedback, reasoning depth can be tuned through a reasoning-effort setting, and reasoning context from earlier messages can be preserved across turns. The repository also documents the two earlier releases in the series: Qwen3.6, which focused on stability and more fluent agentic coding and repository-level reasoning, and Qwen3.5, which introduced a unified vision-language foundation.

More in Models

See all 405
  • mini-AGI

    github.com

    A continual learning language model that trains from scratch on a single 8 GB VRAM GPU, assembling its own architecture and adapting its size to the hardware and data available rather than shipping as a fixed pretrained model.

  • An open-source contrastive-embedding pipeline that maps technical CVE vulnerability descriptions to formal cybersecurity compliance controls in NIST SP 800-53 and CMMC, published with its model generation, fine-tuning and evaluation code.

  • peekaboolean

    github.com

    A small vision-language model, fine-tuned with LoRA on a 500-million-parameter backbone, that answers typed multiple-choice, ordinal-score and yes/no questions about an image with calibrated probabilities, in about 400 milliseconds on a laptop.