Qwen3 235B-A22B
A 235 billion-parameterHow many numbers the model learned during training — the usual rough measure of its size. mixture-of-expertsA model split into parts where only a few run for each word, so it answers faster than its size suggests. model using ~22 billion parameters per token from Qwen Team (Alibaba Cloud). Dense and MoE models with hybrid thinking modes.
- Size
- 235B22B active
- Memory to run
- —smallest, 8K ctx
- Context
- 256Ktokens
- License
- PermissiveApache License 2.0
- Updated
- 18 Sept 2026
Which version to use
- Variant Qwen3 235B A22B Instruct 2507 · Qwen Team (Alibaba Cloud)Tuned to follow instructions and hold a conversation. The usual choice.
Other sizes in Qwen3: Qwen3 0.6B, Qwen3 1.7B, Qwen3 4B, Qwen3 8B, Qwen3 14B, Qwen3 30B-A3B, Qwen3 32B
Where it runs
Best-fitting version on each reference system, at an 8K contextHow much text the model can consider at once, counted in tokens — roughly ¾ of a word each..
Benchmarks
Each benchmarkA fixed set of questions every model is given, so their scores can be compared on the same task. is developer-reported; prompts and settings differ between labs. Bars are relative to the best open result.
Variants & downloads
Each variant is a separate set of weights. Expand one for its downloads: quantizationStoring each of the model’s numbers with fewer bits, so the file is smaller and needs less memory. trades a little quality for a much smaller file, and the memory column adds the working memoryExtra space the model needs while it answers. It grows with the length of the conversation, on top of the file itself. a conversation needs on top.
Variant Qwen3 235B A22B Instruct 2507Permissive
- What it is
- Tuned to follow instructions and hold a conversation. The usual choice.
- Publisher
- Qwen Team (Alibaba Cloud)
- License
- Apache License 2.0 · commercial use allowed
- Released
- 21 Jul 2025
- huggingface
- Qwen/Qwen3-235B-A22B-Instruct-2507
Contributions are not open yet, so there is nothing here from members.
No downloadable artifacts recorded.
Architecture details
- Parameters
- 235B
- Active / token
- 22B
- Architecture
- Mixture of experts
- Layers
- 94
- Attention heads
- 64
- KV heads
- 4
- Head dim
- 128
- KV cache @ 8K (fp16)
- 1.47 GiB
- Max context
- 262,144 tokens
Measured performance
Throughput on specific systems and runtimes, with the source of each measurement.
Community results
Reviews
Lineage
How this model’s variants relate to each other and to other models.
Sources & history
Sources
- Hugging Face Hub (live source, 1 record, 18 Sept 2026)
External identifiers
- huggingface Qwen/Qwen3-235B-A22B-Instruct-2507
Field history
- auto_promotion = {"repo":"Qwen/Qwen3-235B-A22B-Instruct-2507","rule":"first_party_release","evidence":["publisher Qwen is Qwen Team (Alibaba Cloud), developer of Qwen, Qwen Coder, Qwen Math","repo name Qwen3-235B-A22B-Instruct-2507 belongs to family Qwen","config.json: 94 layers, 64 attention heads, 4 KV heads, head dim 128, context 262144","safetensors metadata: 235,093,634,560 parameters, consistent with 235B","128 experts (8 per token); A22B active parameters as named by the publisher","published 2025-07-21 (Hugging Face repository creation date)","variant kind instruct stated by the name"]} — Hugging Face Hub (current)
Timeline
- Official Qwen3 GGUF quantizations published
- Qwen3 released with MoE and hybrid thinking