Qwen3-VL-235B-A22B-Instruct GPU Requirements: VRAM & Cheapest GPU
Qwen3-VL-235B-A22B-Instruct has about 236B parameters. See exactly how much GPU memory it needs at FP16, INT8, and INT4, and the cheapest GPU to run it, with live hourly pricing from 5+ data center partners.
To run Qwen3-VL-235B-A22B-Instruct for inference at FP16, you need roughly 514 GB of VRAM. The cheapest fit on Spheron is 8x A100 80GB at about $6.56/hr. Quantize to INT4 to run it on a smaller, cheaper GPU.
Estimated peak VRAM including weights, activations, and KV cache. Add 10% headroom for production traffic.
- 01$0.82/hr$6.56/hr8× A100 80GBCHEAPESTAmpere · HBM2e
- 02$3.32/hr$6.64/hr2× B300 288GBBlackwell Ultra · HBM3e
- 03$0.91/hr$7.28/hr8× RTX PRO 6000 96GBBlackwell · GDDR7
- 04$2.69/hr$10.76/hr4× B200 192GBBlackwell · HBM3e
- 05$1.49/hr$11.92/hr8× H100 80GBHopper · HBM3
Live pricing aggregated from 5+ data center partners. Per-minute billing, no commitments.
VRAM required to run Qwen3-VL-235B-A22B-Instruct
Estimated peak VRAM at context length 4,096 and batch size 1, including weights, activations, and KV cache. Quantizing to INT8 (Q8) or INT4 (Q4) cuts memory roughly in half and in quarter.
| Precision | Inference | LoRA fine-tune | Full fine-tune |
|---|---|---|---|
| FP16 | 514 GB | 771 GB | 2055 GB |
| INT8 | 257 GB | 385 GB | 1028 GB |
| INT4 | 128 GB | 193 GB | 514 GB |
Cheapest GPU to run Qwen3-VL-235B-A22B-Instruct by precision
Full precision. Best quality, highest memory.
8-bit quantized. ~2x smaller, minimal quality loss.
4-bit quantized. ~4x smaller, runs on smaller GPUs.
Inference vs fine-tuning Qwen3-VL-235B-A22B-Instruct
Inference only holds the model weights plus a KV cache, so it is the cheapest setup. LoRA fine-tuning adds a small adapter and roughly 50% more memory. Full fine-tuning holds gradients and optimizer state on top of the weights, which is about 4x the inference footprint, so it often needs multiple GPUs even when inference fits on one. For Qwen3-VL-235B-A22B-Instruct, an on-demand A100 80GB instance covers inference and LoRA, while a full fine-tune needs several times that memory and often spans multiple GPUs. Check the live GPU pricing for current rates.
Deployment guideDeploy Qwen 3 step by stepHands-on production setup, GPU configs, and benchmarks for Qwen3-VL-235B-A22B-Instruct.Read guideSimilar models
Compare GPU requirements for models in the same class.