Alibaba to open-source Qwen3.8-Flash-Next, first look at architecture powering Qwen4
calabi_and_yau · x · 2026-08-25
Alibaba is about to release Qwen3.8-Flash-Next, an open-weight multimodal MoE model built on the next-generation architecture that will underpin the upcoming Qwen4 family. The move gives developers early access to the architectural improvements behind Qwen4. The poster notes Qwen3.8 27B performs great and expresses excitement for Qwen4 27B.
Related event: Alibaba Teases Qwen3.8-Flash-Next as a Preview of Qwen4 Architecture(6 posts)→
More from Models
- OpenAI's new inference engine boosts throughput up to 4.1x on select models — rohanpaul_ai · 2026-08-25
- OpenAI's Codex lead: next-gen models need more than your laptop, Codex will look primitive — koltregaskes · 2026-08-25
- AI4Bharat's 0.8B IndicOCR Takes On Models 4-5x Its Size, Open Weights — sumanthd17 · 2026-08-25
- Developer Eagerly Anticipates Release of Qwen 3.8 Small Models — chris_j_paxton · 2026-08-25
- AI writing latency stagnant; RL focus may degrade quality — peterwildeford · 2026-08-25
- LLM use is like MCMC: reward-seeking prompts that never backtrack — rickasaurus · 2026-08-25