Qwen 3.5 397B
Qwen 3.5 397B
The big open Qwen: closest an Apache licence has come to a frontier model.
Updated
The read
This is the large sparse model from the previous open Qwen generation, and it remains one of the strongest weights anyone has published under a permissive licence. It reasons well, handles enormous context and is a serious fine-tuning base for anyone with the hardware to train on it. It is a generation old now, and the newer, smaller releases beat it on cost per answer.
Where it shines
- One of the strongest Apache-licensed models ever published
- Huge context window on weights you control
- Proven fine-tuning base with a large adapter ecosystem
Worth knowing first
- A generation behind the current open release
- Needs multiple high-memory GPUs to serve properly
Index score
81.9
Provisional. This score is read from the product's public capability, not from the one-prompt rebuild the fully reviewed entries went through, so treat it as a placing rather than a verdict.
- Rank
- 36 of 73
- Builds
- LLMs
- Category
- Open weights
- Output
- Apache-licensed weights, large sparse model
- Pricing
- Free tier
- Best for
- Fine-tuning a large permissively licensed base model
Scores are our own editorial judgement, weighted craft 55, speed 10, control 15 and value 20.
Also worth a look
Full index- 05GLM-5.3 FlashGLM-5.3 FlashProvisionalA fast, MIT-licensed open model that human voters rank beside the big flagships.Open weightsFree tier89.6
- 06Xiaomi MiMo V2.6 ProMiMo V2.6 ProProvisionalThe strongest large open-weights model on the independent index, under MIT, for under a dollar out.Open weightsFree tier89.4
- 07DeepSeek V4.1 FlashDeepSeek V4.1 FlashProvisionalFrontier-adjacent answers for thirty cents per million input tokens, with the weights under MIT.Open weightsFree tier88.3
