Qwen3.8-2.4T-A95B
Canonical page on the main site: sinoaihub.com/models/qwen38-24t-a95b
Type
model
Definition
Qwen3.8-2.4T-A95B is a Alibaba Cloud (Qwen) model in the Qwen3.8 family: 2.4T-parameter MoE, 95B activated, 512 experts (10 routed + 1 shared per token), 92 layers, Gated DeltaNet + Gated Attention hybrid; 262,144-token context window; open-weight; Qwen3.8-Max License (custom MIT-style: unrestricted use/copy/modify/sell, but products with >100M MAU or >US$20M/month revenue must display the model name; Model-as-a-Service or AI Work Assistant businesses with >US$50M/12-month revenue need a separate license from Qwen) license; released 2026-08-12.
Key facts
- Architecture: 2.4T-parameter MoE, 95B activated, 512 experts (10 routed + 1 shared per token), 92 layers, Gated DeltaNet + Gated Attention hybrid
- Context window: 262,144 tokens
- Weights: open; Qwen3.8-Max License (custom MIT-style: unrestricted use/copy/modify/sell, but products with >100M MAU or >US$20M/month revenue must display the model name; Model-as-a-Service or AI Work Assistant businesses with >US$50M/12-month revenue need a separate license from Qwen) license
- Released: 2026-08-12
- Capabilities: reasoning
Provider
Model Family
Qwen3.8
Release Date
2026-08-12
Status
active
Architecture
2.4T-parameter MoE, 95B activated, 512 experts (10 routed + 1 shared per token), 92 layers, Gated DeltaNet + Gated Attention hybrid
Parameter Information
- total_parameters: 2.4T
- active_parameters: 95B
Context Window
262144
Capabilities
- reasoning: Yes
- vision: No
Open Weight
Yes
License
Qwen3.8-Max License (custom MIT-style: unrestricted use/copy/modify/sell, but products with >100M MAU or >US$20M/month revenue must display the model name; Model-as-a-Service or AI Work Assistant businesses with >US$50M/12-month revenue need a separate license from Qwen)
Self Hosting
Yes
Api Available
No
Official Api
No
Known Limitations
- Text-only input; thinking cannot be disabled; reasoning_effort xhigh/medium/low
- Native context 262,144 tokens, extensible to 1,010,000
- Not the same product as the qwen3.8-max API model (which adds vision/video input, non-thinking mode and 1M default context)
Verification Status
verified
License Changes
No documented license changes on record as of 2026-09-29.
Last Verified
2026-09-20
Source history
No documented source-change events located as of 2026-09-20.
Sources
| evidence_id | source_name | source_url | source_type | published | verified | confidence | conflict |
|---|---|---|---|---|---|---|---|
| src-models-qwen3.8-2.4t-a95b-1 | Hugging Face model card — Qwen3.8-2.4T-A95B | https://huggingface.co/Qwen/Qwen3.8-2.4T-A95B | Model card | — | 2026-09-20 | high | — |
| src-models-qwen3.8-2.4t-a95b-2 | Qwen3.8 repository README | https://github.com/QwenLM/Qwen3.8 | Official documentation | — | 2026-09-20 | high | — |
| src-models-qwen3.8-2.4t-a95b-3 | Qwen3.8-2.4T-A95B license file | https://huggingface.co/Qwen/Qwen3.8-2.4T-A95B/raw/main/LICENSE | Official documentation | — | 2026-09-20 | high | — |