Qwen3-235B-A22B-Thinking-2507
- In Input modalities: Text Out Output modalities: Text
- Reasoning
- Tool use
- Open weights
Prices
3 offerings · prices as listedqwen3-235b-a22b-thinking-2507| Conditions/ 1M tokens | Input | Output |
|---|---|---|
| Standard no longer listed since Oct 11, 2026 | $0.23 | $2.30 |
Prices from Alibaba Cloud Model Studio pricing (opens in a new tab) (checked Oct 11, 2026)
qwen3-235b-a22b-thinking-2507| Conditions/ 1M tokens | Input | Output |
|---|---|---|
| Standard no longer listed since Oct 11, 2026 | ¥2.00 | ¥20.00 |
Prices from 阿里云百炼：模型价格 (opens in a new tab) (checked Oct 8, 2026)
Standard: the price a request with no conditions pays (the headline price elsewhere on the site).
Details
- Max input
- 126.98K
- Structured output
- —
- Licence
- apache-2.0
- Family
- qwen-thinking
- Model ID
Not stated on Alibaba Qwen’s pages: reasoning efforts, reasoning can be off, max thinking budget, built-in tools, knowledge cutoff, aliases.
Run locally
Weights: Qwen/Qwen3-235B-A22B-Thinking-2507- Parameters 235.1B
- Active 22B · 128 experts, 8 per token
- Native precision BF16
- Layers 94
- KV cache per token 188 KB
- Native context 256K (model card)
- Config positions 256K (config.json max_position_embeddings)
- API context 128K (maker’s API)
- Needs
- ≈ 123.6 GB (weights 116.3 + KV cache 1.5 + overhead 5.8)
- Usable
- 21.6 GB (90% of its memory)
- Try
- IQ2_XXS runs (needs system memory), but below 4 bits quality drops noticeably
| Precision | Weights | Needs at 8K | On your hardware | Longest context |
|---|---|---|---|---|
| BF16 (native)maker’s file | 437.9 GB | ≈ 461.3 | ✕Doesn’t fit | — |
| Q8_0 | 232.6 GB | ≈ 245.7 | ✕Doesn’t fit | — |
| FP8maker’s file | 220.2 GB | ≈ 232.7 | ✕Doesn’t fit | — |
| Q6_K | 179.8 GB | ≈ 190.3 | ✕Doesn’t fit | — |
| Q5_K_M | 155.7 GB | ≈ 165.0 | ✕Doesn’t fit | — |
| Q5_0 | 150.5 GB | ≈ 159.5 | ✕Doesn’t fit | — |
| Q4_K_M | 133.0 GB | ≈ 141.1 | ✕Doesn’t fit | — |
| Q4_0 | 123.2 GB | ≈ 130.8 | ✕Doesn’t fit | — |
| IQ4_XS | 116.3 GB | ≈ 123.6 | ✕Doesn’t fit | — |
| Q3_K_Mquality drops | 107.0 GB | ≈ 113.8 | ✕Doesn’t fit | — |
| IQ3_XXSquality drops | 83.7 GB | ≈ 89.4 | ✕Doesn’t fit | — |
| Q2_Kquality drops | 82.1 GB | ≈ 87.7 | ✕Doesn’t fit | — |
| IQ2_XXSquality drops | 56.4 GB | ≈ 60.7 | ○Needs system memory 12–18 tok/s | — |
Click a row to choose a precision. “Maker’s file” sizes are the maker’s own files; others are parameters × bits per weight.
Compared with other hardware
Each at the highest precision it runs. Click one for details.
Smallest that runs it (Q4_K_M, 8K): GPU NVIDIA B200 (HGX B200); unified memory Apple M5 Ultra (64-core GPU) · 256 GB.
Run commands
The maker publishes no files at this precision. Choose its native precision or one of its own builds, or quantise it yourself with llama.cpp.
Memory to fine-tune
Batch 1, 2,048-token sequences, gradient checkpointing; LoRA and QLoRA train rank-16 adapters on the attention projections.
Estimates; facts from the maker’s Hugging Face repositories, as of 2026-11-06. How it’s estimated
Cited scores
From third-party leaderboards, not measured by us- LMArena Text 1,400
#98 of 163 listed models · CI 1,394–1,407 · 8,822 votes
- Epoch AI · GPQA Diamond 80.1%
#79 of 168 listed models · ±2.6
- Epoch AI · Mystery Game Puzzles 9%
#67 of 77 listed models · ±2.88
- Epoch AI · OTIS Mock AIME 2024–2025 86.7%
#67 of 159 listed models · ±5.12
- LMArena (Arena), leaderboard-dataset on Hugging Face, licensed CC BY 4.0 · License · as of Oct 28, 2026
- Epoch AI, 'Capabilities & benchmarking'. Published online at epoch.ai. Retrieved from 'https://epoch.ai/benchmarks'. Licensed CC BY 4.0. · License · as of Oct 28 – Nov 4, 2026
- changes we made
Changes
All changes →- Now priced on Alibaba Cloud Model Studio, China Input ¥2.00 · Output ¥20.00
- Context window 256K → 131.07K
- Now priced on Alibaba Cloud Model Studio Input $0.23 · Output $2.30
- Now priced on Alibaba Cloud Model Studio, International Input $0.23 · Output $2.30
Sources
- Text generation models - Qwen, DeepSeek, GLM | Model Studio - Alibaba Cloud Model Studio - Alibaba Cloud Documentation Center (opens in a new tab) docs.modelstudio.console.alibabacloud.com · checked Oct 10, 2026
- qwen3-235b-a22b-thinking-2507 model (opens in a new tab) alibabacloud.com · checked Oct 6, 2026
- Alibaba Cloud Model Studio: newly released models (opens in a new tab) alibabacloud.com · checked Nov 2, 2026
- Alibaba Cloud Model Studio pricing (opens in a new tab) alibabacloud.com · checked Oct 11, 2026
- qwen3-235b-a22b-thinking-2507 (opens in a new tab) docs.modelstudio.console.alibabacloud.com · checked Oct 9, 2026
- qwen3-235b-a22b-thinking-2507 model (opens in a new tab) alibabacloud.com · checked Oct 8, 2026
- 阿里云百炼：模型价格in Chinese (opens in a new tab) help.aliyun.com · checked Oct 8, 2026
- Qwen/Qwen3-235B-A22B-Thinking-2507 model card (opens in a new tab) huggingface.co · checked Nov 1, 2026
- Qwen/Qwen3-235B-A22B-Thinking-2507 config.json (opens in a new tab) huggingface.co · checked Nov 2, 2026
- Qwen/Qwen3-235B-A22B-Thinking-2507 files (opens in a new tab) huggingface.co · checked Nov 1, 2026
- Qwen/Qwen3-235B-A22B-Thinking-2507-FP8 files (opens in a new tab) huggingface.co · checked Nov 3, 2026
- Qwen/Qwen3-235B-A22B-Thinking-2507 README.md (opens in a new tab) huggingface.co · checked Nov 4, 2026
- Qwen/Qwen3-235B-A22B-Thinking-2507 model card (opens in a new tab) huggingface.co · checked Nov 2, 2026
Alibaba Qwen’s official policies: data use, privacy, terms →
Alibaba Qwen · China · qwen.ai
Related models
Snapshots, the same family and other models from this provider