All news

Alibaba Previews Qwen3.8 Max: 2.4T Parameters, Open Weights Promised

Alibaba's Qwen team put a preview of Qwen3.8 Max live — a 2.4-trillion-parameter flagship it pitches as second only to Fable 5 — with open weights promised and access rolling out through Qoder and Model Studio's new flat-rate token plans.


Alibaba's Qwen team launched a preview of Qwen3.8 Max this week, days after Moonshot shipped Kimi K3 — the second Chinese frontier flagship inside a week. The team describes it as a roughly 2.4-trillion-parameter model that is "one of the most powerful models available today," comparable to the leading frontier systems and, in their own framing, second only to Anthropic's Fable 5.

  • Reported scale: ~2.4T parameters, the successor to Qwen3.7 Max, which shipped only two months ago
  • Alibaba says the model is going open-weight — which would make it by far the largest open-weights release to date
  • No official model card or benchmark table yet; the preview follows the quiet-launch playbook Qwen has used on prior Max releases
  • Access is rolling out through Alibaba's coding platforms (including Qoder) and Model Studio, where new flat-rate token plans put the preview behind a subscription rather than pay-as-you-go metering

The launch continues an aggressive 2026 cadence for Qwen — 3.5, 3.6 and 3.7 all shipped this year, with the Max line moving from open-source leader to closed flagship and now, apparently, back toward open weights. Qwen3.8 Max is joining the arena roster: it runs the same one-shot, single-file briefs as every other model here, so you can vote on it blind against Claude, GPT, Kimi K3 and the rest of the field.

More AI news in Polish at nowosci.ai