Alibaba Open-Sources Qwen3.8-Flash, Cutting Costs by 90%
Alibaba has open-sourced Qwen3.8-Flash, a 125B-parameter MoE model activating only 6B per token, cutting training costs by about 90% while reportedly beating Claude Opus 4.6 on key tasks and supporting up to 1M-token context.
2026-08-28 ~ 2026-08-28 · 3 related posts
- Episode 1: Alibaba's Qwen Team Announces Qwen3.8-Flash-Next as a Preview of Qwen4 Architecture(2026-08-25, 13 posts)
- Episode 2: Alibaba Open-Sources Qwen3.8-Flash-Next, an Early Preview of the Qwen4 Architecture(2026-08-26, 26 posts)
- Episode 3: Qwen3.8-Flash runs locally, tops Claude Opus on SWE-bench Pro(2026-08-26, 5 posts)
- Episode 4: Qwen3.8-Flash-Next FP8 Quantized Model Released on Hugging Face(2026-08-27, 2 posts)
- Episode 5: Qwen 3.8-Next Released with Detailed Technical Report(2026-08-27, 2 posts)
- Episode 6: Alibaba Open-Sources Qwen3.8-Flash, Cutting Costs by 90%(2026-08-28, 3 posts)
- Qwen3.8-Flash-Next Released: A Free AI Rivaling Billion-Dollar Giants — Two Minute Papers · 2026-08-28
- Alibaba Open Sources Qwen3.8-Flash: Undercuts DeepSeek, Runs 1M Context on 4090 — 量子位 · 2026-08-28
- Alibaba's Qwen3.8-Flash activates only 6B of 125B params, cuts costs by 90% — 大模型之路 · 2026-08-28