Qwen3 235B A22B Thinking 2507 — Hardware Requirements & GPU Compatibility
Chat
Qwen3 235B A22B Thinking 2507 is the reasoning and chain-of-thought variant of Alibaba's largest Qwen3 mixture-of-experts model, updated in July 2025. With 235 billion total parameters and about 22 billion active per forward pass, it represents the pinnacle of Qwen3's reasoning capabilities. This model excels at complex multi-step problems, mathematical reasoning, code analysis, and tasks requiring deep logical thinking. It demands serious hardware to run locally, but for users with multi-GPU setups, it offers reasoning performance that rivals the best proprietary models while keeping all computation on your own machines.
56.8K downloads
406 likesSpecifications
- Publisher
- Alibaba
- Family
- Qwen 3
- Parameters
- 235.1B
- Architecture
- Qwen3MoeForCausalLM
- Context Length
- 262,144 tokens
- Vocabulary Size
- 151,936
- Release Date
- 2025-07-25
- License
- Apache 2.0
Get Started
HuggingFace
How Much VRAM Does Qwen3 235B A22B Thinking 2507 Need?
Select a quantization to see compatible GPUs below.
| Quantization | Bits | VRAM | + Context | File Size | Quality |
|---|---|---|---|---|---|
| Q2_K | 3.40 | 100.4 GB | 125.5 GB | 99.91 GB | 2-bit quantization with K-quant improvements |
| Q3_K_S | 3.50 | 103.3 GB | 128.4 GB | 102.85 GB | 3-bit small quantization |
| Q3_K_M | 3.90 | 115.1 GB | 140.1 GB | 114.61 GB | 3-bit medium quantization |
| Q4_0 | 4.00 | 118.0 GB | 143.1 GB | 117.55 GB | 4-bit legacy quantization |
| Q4_K_M | 4.80 | 141.6 GB | 166.6 GB | 141.06 GB | 4-bit medium quantization — most popular sweet spot |
| Q5_K_M | 5.70 | 168 GB | 193.0 GB | 167.50 GB | 5-bit medium quantization — good quality/size tradeoff |
| Q6_K | 6.60 | 194.4 GB | 219.5 GB | 193.95 GB | 6-bit quantization, very good quality |
| Q8_0 | 8.00 | 235.6 GB | 260.6 GB | 235.09 GB | 8-bit quantization, near-lossless |
Which GPUs Can Run Qwen3 235B A22B Thinking 2507?
Q4_K_M · 141.6 GBQwen3 235B A22B Thinking 2507 (Q4_K_M) requires 141.6 GB of VRAM to load the model weights. For comfortable inference with headroom for KV cache and system overhead, 185+ GB is recommended. Using the full 262K context window can add up to 25.0 GB, bringing total usage to 166.6 GB. No single GPU has enough memory — multi-GPU or cluster setups are needed.
Which Devices Can Run Qwen3 235B A22B Thinking 2507?
Q4_K_M · 141.6 GB6 devices with unified memory can run Qwen3 235B A22B Thinking 2507, including NVIDIA DGX H100, NVIDIA DGX A100 640GB, Mac Studio (M3 Ultra, 256GB).
Runs great
— Plenty of headroom
Decent
— Enough memory, may be tightWhere to Download Qwen3 235B A22B Thinking 2507
Community quantizations of this model — GGUF for llama.cpp, Ollama, and LM Studio, plus AWQ/MLX variants where available.
Benchmarks
Benchmark details →Related Models
Frequently Asked Questions
- How much VRAM does Qwen3 235B A22B Thinking 2507 need?
Qwen3 235B A22B Thinking 2507 requires 141.6 GB of VRAM at Q4_K_M, or 470.7 GB at BF16. Full 262K context adds up to 25.0 GB (166.6 GB total).
VRAM = Weights + KV Cache + Overhead
Weights = 235.1B × 4.8 bits ÷ 8 = 141.1 GB
KV Cache + Overhead ≈ 0.5 GB (at 2K context + ~0.3 GB framework)
KV Cache + Overhead ≈ 25.5 GB (at full 262K context)
VRAM usage by quantization
Q4_K_M141.6 GBQ4_K_M + full context166.6 GB- Can NVIDIA GeForce RTX 5090 run Qwen3 235B A22B Thinking 2507?
No — Qwen3 235B A22B Thinking 2507 requires at least 71.0 GB at IQ2_XS, which exceeds the NVIDIA GeForce RTX 5090's 32 GB of VRAM.
- What's the best quantization for Qwen3 235B A22B Thinking 2507?
For Qwen3 235B A22B Thinking 2507, Q4_K_M (141.6 GB) offers the best balance of quality and VRAM usage. Q5_K_S (162.1 GB) provides better quality if you have the VRAM. The smallest option is IQ2_XS at 71.0 GB.
VRAM requirement by quantization
IQ2_XS71.0 GBQ2_K100.4 GBQ3_K_L121.0 GBQ4_K_M ★141.6 GBQ5_K_S162.1 GBBF16470.7 GB★ Recommended — best balance of quality and VRAM usage.
- Can I run Qwen3 235B A22B Thinking 2507 on a Mac?
Qwen3 235B A22B Thinking 2507 requires at least 71.0 GB at IQ2_XS, which exceeds the unified memory of most consumer Macs. You would need a Mac Studio or Mac Pro with a high-memory configuration.
- Can I run Qwen3 235B A22B Thinking 2507 locally?
Yes — Qwen3 235B A22B Thinking 2507 can run locally on consumer hardware. At Q4_K_M quantization it needs 141.6 GB of VRAM. Popular tools include Ollama, LM Studio, and llama.cpp.
- How fast is Qwen3 235B A22B Thinking 2507?
At Q4_K_M, Qwen3 235B A22B Thinking 2507 can reach ~31 tok/s on AMD Instinct MI350X. Speed depends mainly on GPU memory bandwidth. Real-world results typically within ±20%.
tok/s = (bandwidth GB/s ÷ model GB) × efficiency
Example: NVIDIA B200 → 8000 ÷ 141.6 × 0.65 = ~37 tok/s
Estimated speed at Q4_K_M (141.6 GB)
~37 tok/s~37 tok/s~31 tok/sReal-world results typically within ±20%. Speed depends on batch size, quantization kernel, and software stack.
- What's the download size of Qwen3 235B A22B Thinking 2507?
At Q4_K_M, the download is about 141.06 GB. The full-precision BF16 version is 470.19 GB. The smallest option (IQ2_XS) is 70.53 GB.
- Which GPUs can run Qwen3 235B A22B Thinking 2507?
No single consumer GPU has enough VRAM to run Qwen3 235B A22B Thinking 2507 at Q4_K_M (141.6 GB). Multi-GPU or professional hardware is required.
- Which devices can run Qwen3 235B A22B Thinking 2507?
6 devices with unified memory can run Qwen3 235B A22B Thinking 2507 at Q4_K_M (141.6 GB), including Mac Pro M2 Ultra (192 GB), Mac Studio (M3 Ultra, 256GB), Mac Studio (M3 Ultra, 512GB), Mac Studio M2 Ultra (192 GB). Apple Silicon Macs use unified memory shared between CPU and GPU, making them well-suited for local LLM inference.