Qwen3-30B-A3B-Instruct-2507
- In Input modalities: Text Out Output modalities: Text
- Tool use
- Structured output
- Open weights
Prices
3 offerings · prices as listedqwen3-30b-a3b-instruct-2507| Conditions/ 1M tokens | Input | Output |
|---|---|---|
| Standard no longer listed since Oct 11, 2026 | $0.20 | $0.80 |
Prices from qwen3-30b-a3b-instruct-2507 model (opens in a new tab) (checked Oct 6, 2026)
qwen3-30b-a3b-instruct-2507| Conditions/ 1M tokens | Input | Output |
|---|---|---|
| Standard no longer listed since Oct 11, 2026 | $0.108 | $0.431 |
Prices from Alibaba Cloud Model Studio pricing (opens in a new tab) (checked Oct 11, 2026)
qwen3-30b-a3b-instruct-2507| Conditions/ 1M tokens | Input | Output |
|---|---|---|
| Standard no longer listed since Oct 11, 2026 | ¥0.75 | ¥3.00 |
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
- Reasoning
- —
- Licence
- Apache 2.0
- Family
- qwen
- Model ID
Not stated on Alibaba Qwen’s pages: max input, reasoning efforts, max thinking budget, built-in tools, knowledge cutoff, aliases.
Run locally
Weights: Qwen/Qwen3-30B-A3B-Instruct-2507- Parameters 30.5B
- Active 3.3B · 128 experts, 8 per token
- Native precision BF16
- Layers 48
- KV cache per token 96 KB
- Native context 256K (model card)
- Config positions 256K (config.json max_position_embeddings)
- API context 128K (maker’s API)
- Needs
- ≈ 18.9 GB (weights 17.3 + KV cache 0.8 + overhead 0.9)
- Usable
- 21.6 GB (90% of its memory)
- Generation speed
- 179–269 tok/s (reads the active 3.3B per token)
- Longest context
- 37K (longer doesn’t fit)
- First token
- about 0.11–0.16 s for a 2K prompt
| Precision | Weights | Needs at 8K | On your hardware | Longest context |
|---|---|---|---|---|
| BF16 (native)maker’s file | 56.9 GB | ≈ 60.5 | ○Needs system memory 12–17 tok/s | — |
| Q8_0 | 30.2 GB | ≈ 32.5 | ○Needs system memory 30–45 tok/s | — |
| FP8maker’s file | 29.0 GB | ≈ 31.2 | ○Needs system memory 34–51 tok/s | — |
| Q6_K | 23.4 GB | ≈ 25.3 | ○Needs system memory 69–103 tok/s | — |
| Q5_K_M | 20.2 GB | ≈ 22.0 | ○Needs system memory 177–265 tok/s | — |
| Q5_0 | 19.5 GB | ≈ 21.3 | ●Runs well 164–246 tok/s | 11K |
| Q4_K_M | 17.3 GB | ≈ 18.9 | ●Runs well 179–269 tok/s | 37K |
| Q4_0 | 16.0 GB | ≈ 17.5 | ●Runs well 189–284 tok/s | 51K |
| IQ4_XS | 15.1 GB | ≈ 16.6 | ●Runs well 197–295 tok/s | 61K |
| Q3_K_Mquality drops | 13.9 GB | ≈ 15.3 | ●Runs well 208–313 tok/s | 75K |
| IQ3_XXSquality drops | 10.9 GB | ≈ 12.2 | ●Runs well 244–366 tok/s | 109K |
| Q2_Kquality drops | 10.7 GB | ≈ 11.9 | ●Runs well 247–370 tok/s | 111K |
| IQ2_XXSquality drops | 7.3 GB | ≈ 8.6 | ●Runs well 305–457 tok/s | 256K |
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 GeForce RTX 4090; unified memory Apple M6 (24GB or 32GB) · 32 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 1,382
#113 of 163 listed models · CI 1,377–1,387 · 23,312 votes
- Epoch AI · GPQA Diamond 55.6%
#118 of 168 listed models · ±2.88
- Epoch AI · OTIS Mock AIME 2024–2025 62.2%
#101 of 159 listed models · ±6.51
- LMArena Text · Chinese 1,433
#95 of 161 listed models · CI 1,418–1,447 · 1,752 votes
- 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 Nov 4, 2026
- changes we made
Changes
All changes →- Now priced on Alibaba Cloud Model Studio, China Input ¥0.75 · Output ¥3.00
- Context window 256K → 131.07K
- Now priced on Alibaba Cloud Model Studio Input $0.108 · Output $0.431
- Now priced on Alibaba Cloud Model Studio, International Input $0.20 · Output $0.80
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
- qwen3-30b-a3b-instruct-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
- qwen3-30b-a3b-instruct-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
- Alibaba Cloud Model Studio pricing (opens in a new tab) alibabacloud.com · checked Oct 11, 2026
- Qwen/Qwen3-30B-A3B-Instruct-2507 model card (opens in a new tab) huggingface.co · checked Nov 2, 2026
- Qwen/Qwen3-30B-A3B-Instruct-2507 config.json (opens in a new tab) huggingface.co · checked Oct 31, 2026
- Qwen/Qwen3-30B-A3B-Instruct-2507 files (opens in a new tab) huggingface.co · checked Nov 1, 2026
- Qwen/Qwen3-30B-A3B-Instruct-2507-FP8 files (opens in a new tab) huggingface.co · checked Oct 31, 2026
- Qwen/Qwen3-30B-A3B-Instruct-2507 README.md (opens in a new tab) huggingface.co · checked Nov 4, 2026
- models.dev: alibaba/qwen3-30b-a3b-instruct-2507 (community data, MIT) (opens in a new tab) github.com · checked Oct 22, 2026
Alibaba Qwen’s official policies: data use, privacy, terms →
Some values on this page are still waiting to be checked against official pages: a filled dot is verified, a hollow one is not yet (with or without a source on file). How we verify data
Alibaba Qwen · China · qwen.ai
Related models
Snapshots, the same family and other models from this provider