Can NVIDIA GeForce RTX 5090 run Qwen3 235B A22B Instruct 2507?

No — Qwen3 235B A22B Instruct 2507 requires at least 71 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 Instruct 2507?

For Qwen3 235B A22B Instruct 2507, Q4_K_M (141.5 GB) offers the best balance of quality and VRAM usage. Q4_K_L (144.4 GB) provides better quality if you have the VRAM. The smallest option is IQ2_XS at 71 GB.

Can I run Qwen3 235B A22B Instruct 2507 on a Mac?

Qwen3 235B A22B Instruct 2507 requires at least 71 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 Instruct 2507 locally?

Yes — Qwen3 235B A22B Instruct 2507 can run locally on consumer hardware. At Q4_K_M quantization it needs 141.5 GB of VRAM. Popular tools include Ollama, LM Studio, and llama.cpp.

How fast is Qwen3 235B A22B Instruct 2507?

At Q4_K_M, Qwen3 235B A22B Instruct 2507 can reach ~21 tok/s on AMD Instinct MI300X. Speed depends mainly on GPU memory bandwidth. Real-world results typically within ±20%.

What's the download size of Qwen3 235B A22B Instruct 2507?

At Q4_K_M, the download is about 141.00 GB. The full-precision Q8_0 version is 235.00 GB. The smallest option (IQ2_XS) is 70.50 GB.

Alibaba·Qwen·Qwen3MoeForCausalLM

Qwen3 235B A22B Instruct 2507 — Hardware Requirements & GPU Compatibility

Chat

Qwen3 235B A22B Instruct 2507 is Alibaba's flagship instruction-tuned model from the July 2025 update, featuring 235 billion total parameters with approximately 22 billion active during inference. As the largest instruct model in the Qwen3 lineup, it delivers top-tier conversational quality, knowledge depth, and instruction following. Despite its massive total parameter count, the MoE architecture keeps active compute manageable. Running this model locally still requires substantial hardware, typically multi-GPU setups with 48 GB or more of total VRAM, but the 2507 refresh makes it one of the most capable open-weight models available for users with high-end local infrastructure.

166.8K downloads 765 likesSep 2025262K context

Specifications

Publisher: Alibaba
Family: Qwen
Parameters: 235B
Architecture: Qwen3MoeForCausalLM
Context Length: 262,144 tokens
Vocabulary Size: 151,936
Release Date: 2025-09-17
License: Apache 2.0

Get Started

HuggingFace

Qwen/Qwen3-235B-A22B-Instruct-2507

How Much VRAM Does Qwen3 235B A22B Instruct 2507 Need?

Select a quantization to see compatible GPUs below.

Quantization	Bits	VRAM	+ Context	File Size	Quality
IQ2_XS	2.40	71 GB	96.0 GB	70.50 GB	Importance-weighted 2-bit, extra small
IQ2_S	2.50	73.9 GB	99.0 GB	73.44 GB	Importance-weighted 2-bit, small
IQ2_M	2.70	79.8 GB	104.8 GB	79.31 GB	Importance-weighted 2-bit, medium
IQ3_XXS	3.10	91.6 GB	116.6 GB	91.06 GB	Importance-weighted 3-bit
IQ3_XS	3.30	97.4 GB	122.5 GB	96.94 GB	Importance-weighted 3-bit, extra small
Q2_K	3.40	100.4 GB	125.4 GB	99.88 GB	2-bit quantization with K-quant improvements
Q3_K_S	3.50	103.3 GB	128.3 GB	102.81 GB	3-bit small quantization
IQ3_M	3.60	106.3 GB	131.3 GB	105.75 GB	Importance-weighted 3-bit, medium
Q3_K_M	3.90	115.1 GB	140.1 GB	114.56 GB	3-bit medium quantization
Q4_0	4.00	118 GB	143.0 GB	117.50 GB	4-bit legacy quantization
Q3_K_L	4.10	120.9 GB	146.0 GB	120.44 GB	3-bit large quantization
IQ4_XS	4.30	126.8 GB	151.8 GB	126.31 GB	Importance-weighted 4-bit, compact
IQ4_NL	4.50	132.7 GB	157.7 GB	132.19 GB	Importance-weighted 4-bit, non-linear
Q4_1	4.50	132.7 GB	157.7 GB	132.19 GB	4-bit legacy quantization with offset
Q4_K_S	4.50	132.7 GB	157.7 GB	132.19 GB	4-bit small quantization
Q4_K_M	4.80	141.5 GB	166.5 GB	141.00 GB	4-bit medium quantization — most popular sweet spot
Q4_K_L	4.90	144.4 GB	169.5 GB	143.94 GB	4-bit large quantization
Q5_K_S	5.50	162.1 GB	187.1 GB	161.56 GB	5-bit small quantization
Q5_K_M	5.70	167.9 GB	193.0 GB	167.44 GB	5-bit medium quantization — good quality/size tradeoff
Q6_K	6.60	194.4 GB	219.4 GB	193.88 GB	6-bit quantization, very good quality
Q8_0	8.00	235.5 GB	260.5 GB	235.00 GB	8-bit quantization, near-lossless

Which GPUs Can Run Qwen3 235B A22B Instruct 2507?

Q4_K_M · 141.5 GB

Show professional

Qwen3 235B A22B Instruct 2507 (Q4_K_M) requires 141.5 GB of VRAM to load the model weights. For comfortable inference with headroom for KV cache and system overhead, 184+ GB is recommended. Using the full 262K context window can add up to 25.0 GB, bringing total usage to 166.5 GB. No single GPU has enough memory — multi-GPU or cluster setups are needed.

Which Devices Can Run Qwen3 235B A22B Instruct 2507?

Q4_K_M · 141.5 GB

4 devices with unified memory can run Qwen3 235B A22B Instruct 2507, including NVIDIA DGX H100, NVIDIA DGX A100 640GB, Mac Pro M2 Ultra (192 GB).

Runs great

— Plenty of headroom

NVIDIA DGX H100~123 tok/s NVIDIA DGX A100 640GB~75 tok/s

Decent

— Enough memory, may be tight

Mac Pro M2 Ultra (192 GB)~4 tok/s Mac Studio M2 Ultra (192 GB)~4 tok/s

Related Models

Derivatives (5)

Qwen Qwen3 235B A22B Instruct 2507 GGUF Qwen3 235B A22B Instruct 2507 GGUF Qwen3 235B A22B Instruct 2507 GGUF Qwen3 235B A22B Instruct 2507 GGUF Qwen3 235B A22B Instruct 2507 GGUF

Frequently Asked Questions

How much VRAM does Qwen3 235B A22B Instruct 2507 need?: Qwen3 235B A22B Instruct 2507 requires 141.5 GB of VRAM at Q4_K_M, or 235.5 GB at Q8_0. Full 262K context adds up to 25.0 GB (166.5 GB total).
VRAM = Weights + KV Cache + Overhead
Weights = 235B × 4.8 bits ÷ 8 = 141 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_M
141.5 GB
Q4_K_M + full context
166.5 GB
Learn more about VRAM estimation →
Can NVIDIA GeForce RTX 5090 run Qwen3 235B A22B Instruct 2507?: No — Qwen3 235B A22B Instruct 2507 requires at least 71 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 Instruct 2507?: For Qwen3 235B A22B Instruct 2507, Q4_K_M (141.5 GB) offers the best balance of quality and VRAM usage. Q4_K_L (144.4 GB) provides better quality if you have the VRAM. The smallest option is IQ2_XS at 71 GB.
VRAM requirement by quantization
IQ2_XS
71.0 GB~57%
Q2_K
100.4 GB~75%
Q3_K_L
120.9 GB~86%
Q4_K_M ★
141.5 GB~89%
Q4_K_L
144.4 GB~90%
Q8_0
235.5 GB~99%
★ Recommended — best balance of quality and VRAM usage.
Learn more about quantization →
Can I run Qwen3 235B A22B Instruct 2507 on a Mac?: Qwen3 235B A22B Instruct 2507 requires at least 71 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 Instruct 2507 locally?: Yes — Qwen3 235B A22B Instruct 2507 can run locally on consumer hardware. At Q4_K_M quantization it needs 141.5 GB of VRAM. Popular tools include Ollama, LM Studio, and llama.cpp.
How fast is Qwen3 235B A22B Instruct 2507?: At Q4_K_M, Qwen3 235B A22B Instruct 2507 can reach ~21 tok/s on AMD Instinct MI300X. Speed depends mainly on GPU memory bandwidth. Real-world results typically within ±20%.
tok/s = (bandwidth GB/s ÷ model GB) × efficiency
Example: AMD Instinct MI300X → 5300 ÷ 141.5 × 0.55 = ~21 tok/s
Estimated speed at Q4_K_M (141.5 GB)
AMD Instinct MI300X
~21 tok/s
Real-world results typically within ±20%. Speed depends on batch size, quantization kernel, and software stack.
Learn more about tok/s estimation →
What's the download size of Qwen3 235B A22B Instruct 2507?: At Q4_K_M, the download is about 141.00 GB. The full-precision Q8_0 version is 235.00 GB. The smallest option (IQ2_XS) is 70.50 GB.