Qwen3.8 launching, going open-weight soon — 2.4T params, 'second only to Fable 5'

Source

Qwen3.8 launching, going open-weight soon — 2.4T params, 'second only to Fable 5'

· @Alibaba_Qwen

Source

Qwen3.8 is launching and going open-weight soon!

With a massive 2.4T parameters, this model is continuously evolving. We believe it's one of the most powerful model available today, compatible to leading frontier AI models, second only to Fable 5.

You don't have to wait to test it. Just now, the Qwen3.8-Max-Preview made its debut on Alibaba's Token Plan, Qoder, and QoderWork. Be among the very first to try it out.

— @Alibaba_Qwen, 2026-07-19T08:29:31Z. Notable reply (Ruben Hassid): "'just behind fable' is the new benchmark... somehow i see it as a win for Anthropic and a massive lost for OpenAI - it's been 6+ months OpenAI is no longer the guy to beat."

Context: 48 hours after Moonshot's Kimi K3 (2.8T, open-weight July 27). Second frontier-class Chinese open-weight announcement in one week.

8 events
DateTypeEvent
Jul 19 SourceDean W. Ball on Kimi K3 — genuinely frontier, not distillation, token-hungry, 'not obviously cheap to run' Jul 18 TweetSemiAnalysis: K3's linear attention is bullish, not bearish, for NVIDIA/HBM/networking Jul 18 Desk NoteAI-semi derate: the tape says sector derate not election lull — turn signature armed Jul 17 PickSemafor — Kimi K3 threatens AI business models, not AI compute Jul 17 ArticleThe K3 Scare Is Not the R1 Scare Jul 22 PickAlphabet Q2: $44.9B quarterly capex, $180-190B FY guide, 2027 'significantly increases' Jul 13 Redditr/stocks — TSMC reports 68% surge in June revenue Jul 28 ArticleFifteen of twenty broke together

Tweets, videos, filings, articles, and source pages tied to this read.

1
X@Alibaba_Qwen tweet