Qwen3 235B-A22B Instruct 2507 on RTX 5060 Ti 16GB (16GB) — VRAM, KV Cache & Throughput

Qwen3 235B-A22B Instruct 2507 (235.09B total, 22B active, Sparse MoE + GQA) uses 132.39 GiB of 4-bit weights (Q4_K_M), requiring 134.42 GiB total at 8K context and 138.83 GiB at 32K on RTX 5060 Ti 16GB (~4.2 tok/s decode).

Interactive Sizing Workbench

Full Compatibility Report →

Open-Weight Model235.09B / 22B

Weights & QuantizationQ4_K_M · 132.39 GiB

Context Window8,192 (8K)

Target Hardware

16 GB · 448 GB/s

GPU Count (Tensor Parallel)

Concurrent Batch Size

KV Cache Precision

Inference Engine

System RAM Offload (CPU)0%

Exceeds Available GPU Memory

Measured GGUF shard size HF config.json ↗

Memory Allocation Breakdown (GiB) 134.42 GiB (144.33 GB) / 16 GiB (840%)

Model Weights: 132.39 GiB KV Cache & State: 1.47 GiB Runtime & Activation Buffers: 0.56 GiB System RAM Offload: 0 GiB

Required GPU Memory

134.42 GiB

144.33 GB (decimal)

Usable Hardware Memory

15.2 GiB

16 GB (physical)

Safe Headroom

-119.22 GiB

840% utilization

Est. Decode Throughput

~4.2 tok/s

22B active params

Est. Prefill Throughput

~650 tok/s

4K Prompt TTFT: ~6302 ms

Est. Cost / 1M Tokens

$171.30

Sparse MoE + GQA · Apache-2.0

Recommended Hardware for Current Configuration

Best Single GPU

Instinct MI300X 192GB (192GB)

~$15,000

Best Value / Used GPU

2x A100 80GB SXM

160GB · ~$24,000

Apple Unified Memory

Mac Studio M4 Ultra 256GB

~$5,999

Cloud From

DigitalOcean · $2.59 / hr

1x Instinct MI300X 192GB

Production Launch Commands

llama.cpp (GGUF)

llama-server -hf unsloth/Qwen3-235B-A22B-Instruct-2507-GGUF:Q4_K_M \
  --ctx-size 8192 -ngl 999 --flash-attn

vLLM

vllm serve Qwen/Qwen3-235B-A22B-Instruct-2507 \
  --max-model-len 8192 --gpu-memory-utilization 0.92

Live Token Speed Simulator (TTFT ~6302ms · ~4.2 tok/s)

[Qwen3 235B-A22B Instruct 2507 · 134.42 GiB] allocated on target GPU. Prefill completed in TTFT window; streaming autoregressive tokens at memory-bandwidth-bound rate...

Cloud GPU Comparison & Rent-vs-Buy TCO

Verified public rates across RunPod, Vast.ai, Vultr, and DigitalOcean.

Verified: 2026-11-05

Active usage (hours / day)4 h / day

Local hardware budget ($)

Electricity rate ($/kWh)

Purchase break-even

~21.1 months

Est. local power: $15.96 / mo

Provider	Recommended Instance	Total VRAM	Hourly Rate	Monthly at Selected Hours	Billing Mode	Deploy
RunPod	1x B200 180GB (Community Cloud)	180 GB	$5.98 / hr	$727.17 / mo	Per-second	Deploy GPU ↗
Vast.ai	1x B200 180GB (Verified Host)	180 GB	$7.50 / hr	$912.24 / mo	Spot / On-demand	Deploy GPU ↗
*Vultr ($300 Free Credit)**	4x L40S 48GB (Dedicated Instance · $300 Credit*)	192 GB	$6.68 / hr	$812.77 / mo	Hourly	Deploy GPU ↗
Vultr Cloud GPU	4x L40S 48GB (Dedicated Instance)	192 GB	$6.68 / hr	$812.77 / mo	Hourly	Deploy GPU ↗
DigitalOcean	1x Instinct MI300X 192GB (Dedicated Instance)	192 GB	$2.59 / hr	$314.94 / mo	Per-second	Deploy GPU ↗

Rates reflect public listings on the verification date and fluctuate with regional inventory.

Fine-Tuning VRAM Requirements (QLoRA / LoRA / Full) — Qwen3 235B-A22B Instruct 2507

QLoRA (4b) Exceeds Available GPU Memory

136.55 GiB

Model Weights: 132.39 GiB · Gradients + Optimizer: 3.08 GiB · Act: 1.08 GiB

Recommended GPU: Instinct MI300X 192GB (192GB)

LoRA (16b) Exceeds Available GPU Memory

444.16 GiB

Model Weights: 437.89 GiB · Gradients + Optimizer: 4.84 GiB · Act: 1.43 GiB

Recommended GPU: 4x H200 141GB SXM (564GB)

Full BF16 Exceeds Available GPU Memory

1797.87 GiB

Model Weights: 437.89 GiB · Gradients + Optimizer: 1357.46 GiB · Act: 2.52 GiB

Recommended GPU: Multi-GPU Required

Engineering Sizing Notes & Optimization Playbook

Quick Hardware Verdict

A single RTX 5060 Ti 16GB (16GB) cannot hold 4-bit Qwen3 235B-A22B Instruct 2507 (134.42 GiB at 8K) entirely in VRAM. Use multi-GPU tensor parallelism, offload layers to system RAM, or deploy on a larger cloud GPU.

KV Cache & Runtime Tuning

Qwen3 235B-A22B Instruct 2507 uses Sparse MoE + GQA (235.09B total, 22B active per token). Moving from 8K to 32K context expands FP16 KV cache from 1.47 GiB to 5.88 GiB; enabling 8-bit KV cache (llama.cpp -ctk q8_0 -ctv q8_0 or vLLM --kv-cache-dtype fp8) saves 2.94 GiB at 32K (reducing total VRAM to 135.89 GiB).

VRAM Scaling Across Context Windows (4K to 128K, 4-bit Weights)

Model weights remain constant while attention KV cache grows with context length and batch size.

Quantization Precision & Memory Matrix — Qwen3 235B-A22B Instruct 2507 @ 8K

Quantization Format	Weights (GiB)	KV Cache	Total VRAM (GiB)	Decode Speed	Compatibility (RTX 5060 Ti 16GB)	Recommended GPU	Data Source
Native Weights (BF16 / FP8 / MXFP4 — Lossless) (BF16)	437.89 GiB	1.47 GiB	439.92 GiB (472.36 GB)	~1.3 tok/s	Exceeds Available GPU Memory	Multi-GPU Required	Official checkpoint size
FP8 / GGUF Q8_0 (8-bit — Near-Lossless) (Q8_0)	232.77 GiB	1.47 GiB	234.8 GiB (252.11 GB)	~2.4 tok/s	Exceeds Available GPU Memory	Instinct MI325X 256GB (256GB)	Measured GGUF shard size
GGUF Q6_K (6-bit — High Precision) (Q6_K)	179.76 GiB	1.47 GiB	181.79 GiB (195.2 GB)	~3.1 tok/s	Exceeds Available GPU Memory	Instinct MI325X 256GB (256GB)	Measured GGUF shard size
GGUF Q5_K_M (5-bit — Balanced) (Q5_K_M)	155.36 GiB	1.47 GiB	157.39 GiB (169 GB)	~3.6 tok/s	Exceeds Available GPU Memory	Instinct MI300X 192GB (192GB)	Measured GGUF shard size
Q4_K_M / UD-Q4_K_XL / AWQ (4-bit — Recommended) (Q4_K_M)	132.39 GiB	1.47 GiB	134.42 GiB (144.33 GB)	~4.2 tok/s	Exceeds Available GPU Memory	Instinct MI300X 192GB (192GB)	Measured GGUF shard size
GGUF IQ4_XS (4.25-bit — Compact) (IQ4_XS)	116.86 GiB	1.47 GiB	118.89 GiB (127.66 GB)	~4.8 tok/s	Exceeds Available GPU Memory	Instinct MI300X 192GB (192GB)	Measured GGUF shard size
GGUF Q3_K_M (3-bit — Low VRAM) (Q3_K_M)	104.72 GiB	1.47 GiB	106.75 GiB (114.62 GB)	~5.4 tok/s	Exceeds Available GPU Memory	Instinct MI300X 192GB (192GB)	Measured GGUF shard size
GGUF UD-Q2_K_XL (2-bit — Extreme MoE) (UD-Q2_K_XL)	82.67 GiB	1.47 GiB	84.7 GiB (90.95 GB)	~6.9 tok/s	Exceeds Available GPU Memory	RTX PRO 6000 Blackwell (96GB)	Measured GGUF shard size

Technical Specifications & FAQ

Does Qwen3 235B-A22B Instruct 2507 fit on RTX 5060 Ti 16GB (16GB)?

In 4-bit (Q4_K_M) at 8,192 context, Qwen3 235B-A22B Instruct 2507 requires 134.42 GiB total GPU memory (132.39 GiB weights + 1.47 GiB KV cache + 0.56 GiB runtime buffers) against 15.2 GiB usable on RTX 5060 Ti 16GB, delivering ~4.2 tok/s decode.

How much VRAM does Qwen3 235B-A22B Instruct 2507 need at 32K context?

At 32,768 context tokens, Qwen3 235B-A22B Instruct 2507 (Sparse MoE + GQA) allocates 5.88 GiB for KV cache and state, bringing total 4-bit VRAM to 138.83 GiB (149.07 GB decimal).

Evaluate This Model on Other GPUs

RTX 3060 12GB (12GB)RTX 4060 Ti 16GB (16GB)RTX 4070 Ti Super (16GB)RTX 4080 Super (16GB)RTX 5070 12GB (12GB)RTX 5070 Ti (16GB)RTX 5080 (16GB)RTX 3090 (24GB)RTX 4090 (24GB)RTX 4090 D (24GB)RTX 5090 (32GB)RTX 5090 D (32GB)RTX A6000 (48GB)RTX 6000 Ada (48GB)RTX PRO 6000 Blackwell (96GB)DGX Spark (GB10 128GB) (128GB)L4 24GB (24GB)A40 48GB (48GB)L20 48GB (48GB)L40S 48GB (48GB)A800 80GB SXM (80GB)A100 80GB SXM (80GB)H800 80GB SXM (80GB)H100 80GB SXM (80GB)H20 96GB HBM3 (96GB)H20 141GB HBM3e (141GB)H200 141GB SXM (141GB)B200 180GB (180GB)B300 288GB Blackwell Ultra (288GB)Ascend 910B 64GB (昇腾) (64GB)Ascend 910C 128GB (昇腾) (128GB)Atlas 300I Duo 96GB (310P) (96GB)MLU590 80GB (寒武纪思元) (80GB)DCU K100-AI 64GB (海光深算) (64GB)MetaX C500 64GB (沐曦曦云) (64GB)Kunlunxin P800 64GB (昆仑芯) (64GB)MTT S4000 48GB (摩尔线程) (48GB)Radeon RX 7900 XTX 24GB (24GB)Radeon AI PRO R9700 32GB (32GB)Ryzen AI Max+ 395 128GB (128GB)Instinct MI300X 192GB (192GB)Instinct MI325X 256GB (256GB)Instinct MI355X 288GB (288GB)Arc Pro B60 24GB (24GB)Gaudi 3 128GB HBM2e (128GB)Mac mini M4 Pro 64GB (64GB)MacBook Pro M4 Max 128GB (128GB)Mac Studio M4 Ultra 256GB (256GB)Mac Studio M3 Ultra 512GB (512GB)

Evaluate Other Models on This Hardware

Qwen3.8 Flash Next (125B-A6B)Qwen3.8 27B Qwen3.8 2.4T-A95B Qwen3.6 35B-A3B Qwen3.5 397B-A17B Qwen3.5 122B-A10B Qwen3.5 35B-A3B Qwen3.5 27B Qwen3.5 9B Qwen3.5 4B Qwen3 Coder 480B-A35B Qwen3 Coder Next 80B-A3B Qwen3 Coder 30B-A3B Qwen3 30B-A3B Instruct 2507 Qwen3 32B Qwen3 14B Qwen3 8B Qwen3 4B Instruct 2507 Qwen3 1.7B Qwen3 0.6B Qwen2.5 72B Instruct Qwen2.5 Coder 32B Instruct QwQ 32B Reasoning DeepSeek V4.1 Flash (552B-A16B)DeepSeek V4 Flash DeepSeek V3.2 (685B)DeepSeek R1-0528 671B DeepSeek V3-0324 671B DeepSeek R1 Distill Llama 70B DeepSeek R1 Distill Qwen 32B DeepSeek R1 Distill Qwen 14B DeepSeek R1 Distill Llama 8B gpt-oss-20b gpt-oss-120b Gemma 4 31B IT Gemma 4 26B-A4B IT Gemma 4 12B IT Gemma 4 E4B IT Gemma 4 E2B IT Gemma 3 27B IT Gemma 3 12B IT Gemma 3 4B IT GLM-5.3 Flash (320B-A18B)GLM-5.3 (753B)GLM-5 (744B)GLM-4.5 (355B-A32B)GLM-4.5 Air (110B-A12B)GLM-4 32B-0414 Kimi K3.1 Pro (2.8T-A104B)Kimi K3 (2.8T-A104B)Kimi K2 Instruct (1T-A32B)Kimi-Dev 72B MiniMax M3.1 Flash (428B-A23B)MiniMax M2.5 (229B-A10B)Step-5 Preview (600B-A27B)Step-3.7 Flash (198B-A11B)Step-3.5 Flash (197B-A11B)ERNIE 4.5 300B-A47B (文心)ERNIE 4.5 21B-A3B (文心)Hunyuan A13B Instruct (混元 80B)MiniCPM5 2B (面壁)Llama 4 Maverick 17B-128E (400B)Llama 4 Scout 17B-16E (109B)Llama 3.1 405B Instruct Llama 3.3 70B Instruct Llama 3.1 8B Instruct Llama 3.2 3B Instruct Mistral Large 4 (1T-A52B)Mistral Large 2 123B Mixtral 8x22B Instruct Mixtral 8x7B Instruct Mistral Small 3.2 24B Devstral Small 24B Nemotron 3 Ultra 550B-A55B Nemotron 3 Super 120B-A12B Nemotron 3 Nano 30B-A3B Cohere Command A 111B Phi-4 14B Liquid LFM3 1.7B