Alibaba's Qwen Open-Sources Qwen3.8-Max, a 2.4T-Parameter Flagship MoE Model

Alibaba's Qwen team has officially released and open-sourced its flagship model Qwen3.8-Max (Qwen3.8-2.4T-A95B), a 2.4-trillion-parameter mixture-of-experts model activating roughly 95B parameters per token. It is Qwen's largest model to date, its first multimodal model beyond one trillion parameters, and the first time the team has opened Max-tier weights. The model topped Hugging Face's trending chart shortly after release and received Day-0 support from major cloud platforms and inference engines.

Confirmed

• Parameters and architecture: 2.4T total parameters, 95B active, 512 built-in experts, combining full attention with linear attention (Gated DeltaNet and Gated Attention); an FP8 version is available.
• Capabilities: officially supports 1M context, with tunable reasoning and parallel tool calling, designed for coding and long-horizon agent workflows.
• Multimodal: developer @multimodalart confirmed native vision and agent tool-calling support from the Hugging Face page config; Together AI describes it as Qwen's first multimodal model above 1T parameters.
• License: per @cedricchee, MIT-like but with a commercial revenue threshold for large AI infrastructure or coding-assistant businesses; lenient for most startups.
• Ecosystem: available on Together AI, Baseten and DeepInfra; Day-0 support from SGLang and Nebius Token Factory; DeepInfra pricing as low as $2/M tokens.
• Real-world test: the SGLang team showed the model autonomously running for 4.5 hours to optimize its own SGLang inference serving and verify performance.

Unconfirmed

• Context length descriptions vary: official claims 1M, Together AI lists 256K, DeepInfra lists 262K; no unified official clarification yet.

Why it matters

• This is the first time Qwen has opened Max-tier flagship weights, and the 2.4T scale makes it one of the most closely watched ultra-large open-weight models.
• SGLang's test shows the model can handle hours-long autonomous engineering tasks, marking a breakthrough in long-horizon agentic capability for open models.

2026-08-12 ~ 2026-08-14 · 21 related posts

Full story(5 episodes)→

• Episode 1: Alibaba's Qwen Open-Sources Qwen3.8-Max, a 2.4T-Parameter Flagship MoE Model(2026-08-12, 21 posts)
• Episode 2: Unsloth Shrinks Qwen3.8 by 91% for Local Deployment(2026-08-13, 3 posts)
• Episode 3: vLLM Announces Day-0 Support for Qwen3.8 2.4T Model(2026-08-13, 2 posts)
• Episode 4: Together AI Launches Serverless Inference for Qwen3.8-2.4T-A95B(2026-08-13, 2 posts)
• Episode 5: Alibaba's Qwen3.8-Max Lands on Five Hosted Platforms with Open Weights(2026-08-13, 6 posts)

• Qwen Releases Qwen3.8-2.4T-A95B Model on Hugging Face — de4dee · 2026-08-12
• [source] Qwen3.8-2.4T Model Surfaces on Hugging Face with Agent Capabilities — multimodalart · 2026-08-12
• [source] Qwen's New 2.4T MoE Model Tops Hugging Face Trending — Qwen · 2026-08-12
• Alibaba's Qwen3.8 Goes Open Weight: 2.4T Parameters with 95B Active — Arindam_1729 · 2026-08-12
• Alibaba's Qwen3.8-2.4T-A95B Drops, Autonomously Optimizes Inference Engine — BanghuaZ · 2026-08-12
• Qwen3.8-Max Weights Released: 2.4T Total Parameters — Yuchenj_UW · 2026-08-12
• Qwen3.8-Max Weights Released with Commercial Revenue Caps — cedric_chee · 2026-08-13
• Alibaba releases Qwen3.8 Max: 2.4T-param MoE, 1M context, open for commercial use — AdinaYakup · 2026-08-13
• Qwen Releases Qwen3.8-Max: 2.4T Parameters with 1M Context natively — Scobleizer · 2026-08-13
• [source] Qwen3.8 2.4T MoE FP8 Quantization Hits Hugging Face Trending — Qwen · 2026-08-13
• Alibaba's Qwen Open-Sources Qwen3.8-2.4T-A95B Model — xiaosun86 · 2026-08-13
• Alibaba Releases 2.4T-Parameter Open-Weight Model Qwen3.8-Max — NVIDIAAI · 2026-08-13
• Alibaba Launches 2.4T Parameter Qwen3.8-Max with 1M Context and Multimodal Support — togethercompute · 2026-08-13
• Qwen3.8-2.4T-A95B now live on Together AI: 2.4T params, 256K context for coding and agents — togethercompute · 2026-08-13
• Alibaba Releases Qwen3.8-Max: 2.4T Parameter MoE with Native Multimodal Support — baseten · 2026-08-13
• Alibaba Qwen Open-Sources Flagship Qwen3.8-2.4T-A95B with Trillion Parameters — 智东西 · 2026-08-13
• Qwen Releases 2.4T-Parameter Open-Weight MoE Model Qwen3.8 — rohanpaul_ai · 2026-08-13
• Qwen3.8-2.4T-A95B Lands on DeepInfra at $2/M Tokens — gharik · 2026-08-14
• Alibaba Open-Sources 2.4 Trillion Parameter Qwen3.8 Flagship Model — 大模型之路 · 2026-08-14
• Qwen3.8 Open-Sources 2.4T Parameter MoE Flagship Model — togethercompute · 2026-08-14

1 near-duplicate retellings: AdinaYakup