Models

Qwen3.8-Max

2 synchronized stories carry this tag.

China model census finds safety results rare at launch

SemiAnalysis reviewed 857 identified releases from nine leading Chinese model developers through 15 September 2026. It found a model-specific published safety result for 31 releases, or 3.6%, and an at-launch result for nine, or 1.1%; 813 releases had no safety disclosure in the materials checked.

Why it matters: The census separates public evidence from statements that a model was merely safety-trained or evaluated. That makes disclosure timing and model-specific matching a practical procurement and model-governance question.

Read storyOriginal source · newsletter.semianalysis.com

Qwen releases 2.4T-parameter Qwen3.8 MoE weights under a custom licence

Qwen released Qwen3.8-2.4T-A95B, a text-only mixture-of-experts model with 2.4 trillion total parameters, 95 billion active parameters and 512 experts. The downloadable BF16 and FP8 checkpoints support a native 262,144-token context and thinking-only operation, with documented extension to about one million tokens.

Why it matters: The release puts a very large sparse model into self-hosted reach, but its compute footprint and commercial licence conditions make it materially different from an Apache-licensed open-weight release.

Read storyOriginal source · huggingface.co