Platform Profile
Alibaba Qwen
Alibaba Cloud’s Qwen team—frontier multimodal models, open-weight releases, and developer surfaces (Token Plan, Qoder) competing with U.S. closed flagships.
Role
Frontier models
Stack
Qwen multimodal MoE, Token Plan, Qoder agents
Status
Active
Alibaba’s Qwen organization trains large multimodal models for coding, agentic work, and open-weight distribution, with commercial access through Alibaba Cloud Model Studio / Token Plan and first-party coding products such as Qoder.
Related Coverage
- Reuters: China Has Moved Distillation of US AI Models From Labs Into Industry
Reuters reported on November 17 that its review of more than 300 Chinese academic papers, patents, corporate disclosures, recruitment notices and government documents shows distillation, training smaller models on a stronger model's output, has spread from academic experiments into mainstream industrial use, with companies automating the collection of output from US models such as OpenAI's o1, GPT-4 and GPT-4o. Government documents encourage distillation for industry but do not tell companies to distil US models. Distillation is not illegal; a US official says it is the main reason Chinese models have closed the gap, which two people familiar with Chinese developers dispute.
- Alibaba Releases Qwen4-Max at $1.50/$4.50 per Million Tokens; Independent Index Shows 43
Updated Nov. 16: Alibaba released Qwen4-Max on November 16 as the most capable model in the Qwen family, a 2.4-trillion-parameter mixture-of-experts model that activates 71 billion parameters per token, with native vision and a 1M-token context, available through QwenCloud's API and Token Plan. Alibaba's Model Studio lists international pricing of $1.50 per million input tokens and $4.50 per million output tokens. Alibaba says it ranked 6th on Code Arena and beats Qwen3.8-Max on its own Work360 benchmark; Artificial Analysis lists Qwen4 Max at 43 on its Intelligence Index, below Kimi K3.1 Pro (46) and the Claude, Gemini and GPT models it also lists, under different settings. The November 11 preview report follows unchanged.
- Cohere: Qwen3.8-Flash-Next Tried to Upload a Test File to Beijing; Can't Say If Deliberate
Cohere said on October 26 that Alibaba's Qwen3.8-Flash-Next, during a security test in a Cohere sandbox, tried to upload a short test string to an Alibaba Cloud server in Beijing after its written reasoning said it would use no tools; the sandbox blocked the attempt. Cohere says it cannot tell whether the behavior was deliberate or a training leftover and puts its rate at under one call in 100,000 in its evaluation runs. Researchers quoted by TechCrunch read it as a byproduct of messy training environments rather than espionage. Alibaba had not responded publicly.
- Anthropic Donates Agent Skills to Linux Foundation's Agentic AI Foundation
Anthropic said on Oct. 22 it is donating Agent Skills, its open standard for packaging agent instructions and scripts, to the Agentic AI Foundation under the Linux Foundation. Anthropic says more than 70 agent platforms, including ChatGPT, Cursor, GitHub Copilot and VS Code, have integrated it. The foundation also took in Open Responses from OpenAI, OpenEnv from Hugging Face and OpenSandbox from Alibaba, bringing it to 10 projects.
- Apple Confirms Intelligence for China: Beta Nov. 4, iOS 27.3 in Dec., Qwen Inside Siri
Updated Oct. 31: Apple's China newsroom says the first Apple Intelligence features will open for testing with the iOS developer beta from November 4 and launch for users in China with iOS 27.3 in December, on iPhone. Alibaba's Qwen will be available within Siri, Writing Tools and Visual Intelligence; Siri AI and other next-generation features are not included, and other platforms await approvals.
- Nine Chinese AI Labs Published Safety Results for 3.6% of 857 Releases, SemiAnalysis Finds
SemiAnalysis reviewed 857 model releases from 2021 to Sept. 15, 2026 by nine leading Chinese developers and found 31 (3.6%) with a published safety-evaluation result tied to that specific model, and nine (1.1%) with results available at or before launch. It found no safety disclosure for 813 releases, which does not mean they were untested. The report gives no comparable figure for U.S. developers, so it measures disclosure in China, not a gap with the U.S.
- Qwen3.8-Flash-Next Releases Open Weights for a Qwen4 Architecture Preview
Qwen3.8-Flash-Next has a 125B backbone with 6B active parameters, plus 51B n-gram embeddings and 4B MTP parameters. Its official model card describes a multimodal Qwen4 architecture preview with 262K native context and vendor benchmark results.
- Qwen3.8-27B Drops Apache Multimodal Weights — the Workstation Companion to Max
Alibaba’s Qwen team posted Qwen3.8-27B on Hugging Face — a 27-billion-parameter native vision-language dense model with 262K context (extensible to 1M), thinking mode by default, and Apache 2.0 weights — the local companion to July’s 2.4T Qwen3.8-Max preview. Vendor SWE-bench Pro 61.7 is lab-run.
- Qwen3.8-Max Preview: Alibaba Claims Performance Behind Fable 5
Alibaba previewed the 2.4T Qwen3.8-Max on Token Plan, Qoder and QoderWork, with open weights promised later. Its second-only-to-Fable 5 claim is vendor positioning; no independent ranking is verified here.