Alibaba Qwen 3 family
https://qwenlm.github.io/blog/qwen3/
Alibaba Cloud's Qwen 3 family (2025-04) — open-weights frontier models from 0.5B to 235B-A22B MoE. Apache 2.0 (most sizes) — broadest open-weights family. Includes Qwen 3 32B dense, Qwen 3 30B-A3B MoE, Qwen 3 235B-A22B MoE. Qwen-Coder, Qwen-VL (vision), Qwen-Audio companion variants. Dominant base model for many Chinese AI startups and a top OSS substrate globally.
Active · Verified 2026-10-31 · fresh as of 2026-10-31 · 54 cited facts, 16 with a third-party source · 1 recorded change · Suggest a correction
At a glance
- Tier
- T1
- Latest release
- Qwen3.8 open weights: flagship Qwen3.8-2.4T-A95B + Qwen3.8-27B (Apache-2.0) (2026-08-17) third-party fix
- Pricing
- Free open-weights (Apache 2.0); Alibaba Cloud API pay-per-token ($0.10-$2 per 1M depending on tier); Together / Fireworks hosted vendor fix
- Funding
- Alibaba (BABA) public; ~$200B market cap vendor fix
- Freshness
- 2026-10-31 fix
- Layer
- large-models repo fix
- Licensing
- open-core (estimate) fix
When to use
Optimised for: multilingual (119 languages including Asian + minor European); cost-efficient inference; permissive Apache 2.0 OSS
Anti-fit: China data-residency for regulated Western workloads; less mainstream Western enterprise adoption; Qwen-Max not open-weights
Pros & cons
Pros
Broadest open-weights frontier family (0.5B to 235B); Apache 2.0 (most permissive); 119-language coverage; strong reasoning at small / medium sizes; #1-ranked OSS family on HF for many months 2024-25.
Cons
China-based — regulatory risk for some Western enterprises; multimodal less polished than Gemini 3; Qwen-Max tier (rumored 1T+) is closed.
Claims & capabilities
Qwen 3 family Apache 2.0; 0.5B to 235B-A22B MoE; 119 languages; thinking/non-thinking dual mode; matches DeepSeek R1 on reasoning at smaller sizes
Technical surface
- API surface
- OpenAI-compatible REST (Alibaba Cloud Model Studio); HuggingFace Transformers; vLLM; llama.cpp
- Backend storage
- not applicable — substrate foundation model
- Deployment
- Self-hostable (Apache 2.0 weights) + Alibaba Cloud Model Studio + Together AI + Fireworks + HuggingFace inference + Ollama
- Embedding model
- not applicable — not a memory product
- Multi-tenancy
- not applicable — substrate foundation model
- MCP
- not applicable — substrate foundation model
- A2A
- not applicable — substrate foundation model
- OpenTelemetry
- not applicable — substrate foundation model
Similar systems
Other foundation models (substrate reference) in the catalog, ranked by inbound references.
- Anthropic Claude (foundation models) T1
Claude family of frontier foundation models — Sonnet 4 (workhorse), Opus 4.5 (frontier capability), Haiku 4.5 (fast/cheap). Frontier model used as substrate by Claude Code, Anthropic Memory tool, Cline, Aider, Continue.dev, Cursor, Goose, and dozens of other catalog entries. Memory in Claude is parametric (in weights) — discrete episodic memory lives outside the model in the Claude Memory tool (see Platform-provider memory section).
- OpenAI GPT family (GPT-5 / GPT-4o / o3 / o4) T1
OpenAI's foundation model family — GPT-5 (2025-08 flagship), GPT-4o (multimodal workhorse), o3/o4 reasoning series. Substrate for ChatGPT, the OpenAI API, Microsoft Copilot, Azure OpenAI, and many downstream agents in this catalog (Devin, Codex CLI, Operator, Cursor mode). Parametric memory in weights; discrete user memory ships in ChatGPT Memory feature (separate row).
- Google Gemini 3 family T1
Google DeepMind's Gemini 3 frontier model family — Gemini 3 Pro (top tier), 3 Flash (fast), 3 Nano (on-device). Substrate for Google Workspace AI, Vertex AI Agent Platform, Gemini Code Assist, Project Mariner, and Project Astra. Multimodal-native; 1M-10M context window. Largest context window of any frontier model. Used by Gemini Enterprise Agent Platform Memory Bank (separate row).
- Mistral Large 2 / Mixtral family T1
French frontier-model lab (Paris). Family includes Mistral Large 2 (123B dense, Jul-2024), Mixtral 8x22B (open-weights MoE), Mistral Small 3 (24B, open-weights Jan-2025), Codestral (code specialist). EU-headquartered alternative to US/China model labs; GDPR-friendly data residency. Distinct from the Mistral Le Chat memory feature (already in catalog) — this row covers the model family itself as substrate.
- xAI Grok 4 T1
Elon Musk's xAI frontier model — Grok 4 launched 2025-07. Trained on Colossus (100k+ GPU supercluster in Memphis). Deeply integrated with X/Twitter (real-time tweet data access). Grok 4 Heavy variant for premium tier. No frontier-tier open-weights — Grok 1 (314B MoE) is the only weight release. The x.ai site now brands itself SpaceXSI (page title, checked 2026-10-29).
- Cohere Command R+ / Command A T1
Toronto-based foundation-model lab (Aidan Gomez, ex-Google Brain 'Attention Is All You Need' co-author). Command family is enterprise / RAG-optimised: Command A (Mar-2025, 111B dense, 256k ctx), Command R+ (104B MoE), Command R7B (smallest tier). Announced merger with Aleph Alpha April-2026. Open weights via CC-BY-NC (non-commercial); Cohere-hosted for commercial.
Sources (53 cited facts, 16 third-party)
Each fact links to the page it came from; the badge says who published it (how to read sources).
- What it is Open-weights frontier model family (Qwen 3 dense + MoE; 0.5B–235B) vendor
- What it does & what's distinct Alibaba Cloud's Qwen 3 family (2025-04) — open-weights frontier models from 0.5B to 235B… vendor
- Claims / benefits Qwen 3 family Apache 2.0; 0.5B to 235B-A22B MoE; 119 languages; thinking/non-thinking du… vendor
- Modalities text, code; image (Qwen-VL line); audio (Qwen-Audio); video (Qwen2.5-VL) vendor
- Created 2023-08 (Qwen 1); 2024-02 (Qwen 1.5); 2024-09 (Qwen 2.5); 2025-04 (Qwen 3) vendor
- Latest release Qwen3.8 open weights: flagship Qwen3.8-2.4T-A95B + Qwen3.8-27B (Apache-2.0) (2026-08-17) third-party
- License Apache 2.0 (most Qwen 3 sizes); some larger MoE variants under Qwen License vendor
- GitHub github.com/QwenLM/Qwen3 ~30k+ stars; cumulative QwenLM org repos ~80k+ stars vendor
- Mindshare Top-2 OSS model family globally on HuggingFace downloads (alongside Llama); standard bas… vendor
- Funding Alibaba (BABA) public; ~$200B market cap vendor
- Customers / scale Alibaba Cloud customers; major Chinese AI startups; international OSS deployers via Hugg… vendor
- Pricing Free open-weights (Apache 2.0); Alibaba Cloud API pay-per-token ($0.10-$2 per 1M dependi… vendor
- Compliance China-based; not US-certified (no SOC 2 / HIPAA / FedRAMP at Alibaba Cloud international… vendor
- Data handling Self-hosting recommended for sensitive data; Apache 2.0 weights eliminate hosted depende… vendor
- Deployment Self-hostable (Apache 2.0 weights) + Alibaba Cloud Model Studio + Together AI + Firework… vendor
- HQ Hangzhou, China (Alibaba Cloud) vendor
- Founders / pedigree Alibaba Cloud (DAMO Academy); Qwen team led by Junyang Lin vendor
- Performance MMLU-Pro: 235B-A22B ~80%; AIME 2025: ~85% (235B); HumanEval: ~90%; matches DeepSeek R1 a… vendor
- Code/weights release Apache 2.0 — weights public on HuggingFace + ModelScope vendor
- API surface OpenAI-compatible REST (Alibaba Cloud Model Studio); HuggingFace Transformers; vLLM; lla… vendor
- LLM lock open (Apache 2.0) vendor
- Runtimes any (HuggingFace, vLLM, llama.cpp, MLX, Ollama) vendor
- Validated verticals Chinese enterprise (Alibaba Cloud customers); international OSS deployers; cost-sensitiv… vendor
- Time to running <10min hosted (Alibaba Cloud / Together / Fireworks); <1hr self-host (Ollama / vLLM) vendor
- Anti-fit China data-residency for regulated Western workloads; less mainstream Western enterprise… vendor
- Optimised for multilingual (119 languages including Asian + minor European); cost-efficient inference;… vendor
- Adjacent infrastructure HuggingFace; ModelScope; Together AI; Fireworks; Ollama; vLLM; Alibaba Cloud Model Studio vendor
- Pros Broadest open-weights frontier family (0.5B to 235B); Apache 2.0 (most permissive); 119-… vendor
- Cons China-based — regulatory risk for some Western enterprises; multimodal less polished tha… vendor
- Project & sources qwenlm.github.io/blog/qwen3/ vendor
- obs-langsmith yes third-party
- obs-datadog yes third-party
- obs-weave yes third-party
- obs-langfuse yes third-party
- obs-custom Qwen via OpenAI-compatible endpoints picks up OpenAI-proxy observability tools vendor
- cost-token-budget yes third-party
- cost-prompt-caching yes third-party
- cost-semantic-caching no third-party
- cost-batching yes third-party
- cost-model-routing no third-party
- cost-streaming-only yes third-party
- cost-observability-cost-attribution no third-party
- eval-langsmith-evals yes third-party
- eval-braintrust yes third-party
- eval-weights-and-biases-agent yes third-party
- eval-helicone-evals yes third-party
- eval-custom-test-harness no vendor
- eval-human-loop no vendor
- eval-production-traffic-replay no vendor
- capability-benchmark-sources livecodebench: 90.3% (Qwen3.8-27B, LiveCodeBench v6, 2026-08) https://huggingface.co/Qwe… registry
- architecture-deployment-locus hybrid vendor
- positioning-platform-vs-product both vendor
- taxonomy-three-layer-rollup large-models repo
What changed
Changes to this record's lifecycle, licence, release, pricing and hosting facts, from the catalog's git history. Subscribe (Atom)
Related systems
Referenced by (7)
- Chroma integrates with — Chroma's docs have a Alibaba Qwen 3 family integration page. (verified 2026-10-29)
- Kong AI Gateway integrates with — Kong AI Gateway docs have a DashScope (Qwen) AI-provider reference page. (verified 2026-10-29)
- LangChain (framework) integrates with — LangChain docs have a Qwen chat model integration page. (verified 2026-10-29)
- Langfuse integrates with — Langfuse's integrations directory has a Alibaba Qwen 3 family integration page. (verified 2026-10-29)
- OpenRouter integrates with — OpenRouter lists Alibaba as an inference provider (providers API slug "alibaba"). (verified 2026-10-29)
- Pipecat (Daily) integrates with — Pipecat docs have a Alibaba Qwen 3 family service page. (verified 2026-10-29)
- Vercel AI SDK integrates with — AI SDK docs have an Alibaba official AI SDK provider page. (verified 2026-10-29)