|

Alibaba Qwen Releases Qwen3.8-Max: A 2.4 Trillion Parameter MoE Model and the Most Capable One in the Qwen Family to Date

ByRicardo

Alibaba’s Qwen group has made Qwen3.8-Max broadly out there and confirmed that its open weights ship subsequent week. A second checkpoint, Qwen3.8-27B, can be going open-weights. Qwen3.8-Max is a 2.4-trillion-parameter mixture-of-experts mannequin. It accepts textual content, picture and video as enter and returns textual content.

Is it deployable

Yes, however the deployable floor relies on which artifact you’re making use of.

The hosted API is deployable as we speak by any firm measurement. It is OpenAI- and DashScope-compatible, so integration is a base-URL and model-ID change. The open weights are a distinct matter. At 2.4T whole parameters, the checkpoint is a multi-node datacenter artifact. Alibaba has not disclosed the activated-parameter rely. Serving price subsequently can not but be modeled. Qwen3.8-27B is the checkpoint that matches peculiar on-premise GPU {hardware}.

The printed function set maps cleanly onto 4 industries. Those are software program engineering, authorized and monetary doc evaluation, media and e-commerce operations, and design.

Applications embrace repository-scale coding brokers and long-document data bases. Long-video indexing, structured knowledge extraction and multi-step analysis assistants additionally match.

Interactive explainer


What is Technically Available

The model page lists a 1M-token context window. Maximum enter is 991K tokens, dropping to 983K when pondering is enabled. Maximum output is 131K tokens in each modes, and the most reasoning price range is 262K tokens. Rate limits are 2M tokens per minute and 15K requests per minute.

Pricing is $2.00 per 1M enter tokens and $6.00 per 1M output tokens. Implicit cache reads price $0.25 per 1M tokens. Explicit cache creation is $2.50 and specific cache reads are $0.17 per 1M tokens. Cached enter is eight instances cheaper than recent enter. Prefix stability subsequently drives price greater than immediate size does.

Supported capabilities embrace function calling, structured outputs, batches, prefix completion and fine-tuning. Five built-in instruments ship on the Responses API: code_interpreter, web_search, web_extractor, t2i_search and i2i_search.

https://qwen.ai/weblog?id=qwen3.8

Performance

Alibaba printed a full benchmark desk with this launch. Qwen3.8-Max scores 86.6 on Terminal-Bench 2.1, forward of Claude Opus 4.8 and Claude Fable 5 at 84.6, behind GPT-5.6 Sol (max) at 88.8. It experiences 67.7 on SWE-bench Pro in opposition to Fable 5’s 80.0, and 73.5 on FrontierSWE in opposition to Fable 5’s 88.8. It leads PaperBench at 93.0 and IFBench at 82.8. GPQA Diamond lands at 92.6, up marginally from Qwen3.7-Max’s 92.4. The clearest positive aspects are multimodal and agentic, not reasoning. It tops most imaginative and prescient rows, together with OSWorld-Verified 86.1, Parametric CAD Bench 91.5, and OmniDocBench 1.5 at 92.1. Against its personal predecessor the bounce is massive: DeepSWE 1.1 strikes from 21.6 to 56.6, FrontierSWE from 40.7 to 73.5, JobBench from 31.3 to 53.4. Two caveats belong in any trustworthy learn. The multimodal desk benchmarks in opposition to Qwen3.7-Plus, not Qwen3.7-Max, which flatters the generational delta. And Alibaba’s personal RL scaling curve peaks at 0.725 close to 4,000 coaching environments, then declines to 0.719 and 0.689.

Key Takeaways

  • Qwen3.8-Max is a 2.4T-parameter MoE mannequin with 1M context, now typically out there.
  • Pricing is $2 enter, $6 output and $0.25 cached enter per 1M tokens.
  • Open weights for Qwen3.8-Max and Qwen3.8-27B are promised subsequent week.
  • No benchmark desk, license, or activated-parameter rely has been printed.
  • The 27B checkpoint, not the flagship, is the real looking on-premise deployment path.

Check out the Technical details, API and Qwen Studio. Also, be at liberty to comply with us on Twitter and don’t neglect to be a part of our 150k+ML SubReddit and Subscribe to our Newsletter. Wait! are you on telegram? now you can join us on telegram as well.

Need to companion with us for selling your GitHub Repo OR Hugging Face Page OR Product Release OR Webinar and so on.? Connect with us

The publish Alibaba Qwen Releases Qwen3.8-Max: A 2.4 Trillion Parameter MoE Model and the Most Capable One in the Qwen Family to Date appeared first on MarkTechPost.

Similar Posts