Divyansh Agrawal (independent) · Released October 19, 2026New
Qwen3.5 4B-Qwenputer has 9.36B parameters. In BF16 its weights alone take 17.4 GiB; quantized to 4-bit, about 6.77 GiB. Each 32K-token request adds 1.00 GiB of KV cache. Cheapest way to run it today: 1 × RTX A4000 (16 GB) at $0.072 per hour. It calls tools and has a thinking mode you can switch off.
Size your own setup in the calculator →Divyansh Agrawal (independent)Weights plus the KV cache of one request (BF16 cache) and the model's fixed recurrent state, in GiB. Add 1 to 3 GiB for the inference engine itself.
| Weights format | Weights only | + 8K context | + 32K context | + 128K context | + 256K context |
|---|---|---|---|---|---|
| BF16 / FP16 | 17.4 | 17.7 | 18.5 | 21.5 | 25.5 |
| FP8 / INT8 | 10.5 | 10.8 | 11.6 | 14.6 | 18.6 |
| 4-bit (GGUF Q4_K_M) | 6.77 | 7.09 | 7.84 | 10.8 | 14.8 |
Fewest GPUs that fit the weights plus one 8K-token request. Data-center GPUs use vLLM (90% of memory usable), the others llama.cpp. Several GPUs split the model with tensor or pipeline parallelism.
| GPU | BF16 / FP16 | 8-bit | 4-bit |
|---|---|---|---|
| GeForce RTX 4090 24 GB | 1 | 1 | 1 |
| GeForce RTX 5090 32 GB | 1 | 1 | 1 |
| RTX PRO 6000 Blackwell 96 GB | 1 | 1 | 1 |
| Mac M5 Max 128 GB 96 GB usable | 1 | 1 | 1 |
| Ryzen AI Max+ 395 128 GB 96 GB usable | 1 | 1 | 1 |
| DGX Spark (GB10) 128 GB | 1 | 1 | 1 |
| NVIDIA H100 80 GB | 1 | 1 | 1 |
| NVIDIA H200 141 GB | 1 | 1 | 1 |
| NVIDIA B200 (HGX) 180 GB | 1 | 1 | 1 |
8-bit: FP8 with vLLM (INT8 on GPUs without FP8), Q8_0 with llama.cpp. 4-bit: AWQ / GPTQ with vLLM, Q4_K_M with llama.cpp, both counted at the Q4_K_M size.
For each GPU: the fewest cards that fit Qwen3.5 4B-Qwenputer with one 8K-token request, the most faithful weight format at that count, and the cheapest on-demand price (2026-11-05).
| Setup | Per hour | Weights | Provider | Per month | Rent |
|---|---|---|---|---|---|
| 1 × RTX A4000 16 GB | $0.072 | Q8_0 | Vast.ai | $52.56 | Rent 1 × RTX A4000 on Vast.ai (opens in a new tab) → |
| 1 × GeForce RTX 3090 24 GB | $0.096 | BF16 | Vast.ai | $70.08 | Rent 1 × GeForce RTX 3090 on Vast.ai (opens in a new tab) → |
| 1 × NVIDIA L4 24 GB | $0.104 | FP8 | Vast.ai | $75.92 | Rent 1 × NVIDIA L4 on Vast.ai (opens in a new tab) → |
| 1 × GeForce RTX 4090 24 GB | $0.140 | BF16 | Vast.ai | $102.20 | Rent 1 × GeForce RTX 4090 on Vast.ai (opens in a new tab) → |
| 1 × RTX A5000 24 GB | $0.201 | BF16 | Vast.ai | $146.73 | Rent 1 × RTX A5000 on Vast.ai (opens in a new tab) → |
| 1 × GeForce RTX 5080 16 GB | $0.201 | Q8_0 | Vast.ai | $146.73 | Rent 1 × GeForce RTX 5080 on Vast.ai (opens in a new tab) → |
| 1 × RTX 2000 Ada 16 GB | $0.240 | Q8_0 | RunPod | $175.20 | Rent 1 × RTX 2000 Ada on RunPod (opens in a new tab) → |
| 1 × RTX A4500 20 GB | $0.250 | Q8_0 | RunPod | $182.50 | Rent 1 × RTX A4500 on RunPod (opens in a new tab) → |
On-demand prices from AWS, Google Cloud, Azure, RunPod, Vast.ai, Oracle Cloud, OVHcloud, Scaleway, Verda, Akash and UpCloud, checked every hour. Serving many users needs more KV cache, so more memory: size it in the calculator. Some links are affiliate links: StudioTV may earn a commission, at no extra cost to you.
AWS, Google Cloud and UpCloud prices: Spare Cores (opens in a new tab), CC BY-SA 4.0 (opens in a new tab), shared under the same license.
Reasoning. Thinking mode you can switch on or off. With vLLM or SGLang, set it per request: "chat_template_kwargs": {"enable_thinking": false} turns it off, true on. Reasoning tokens take context and KV cache like any other token: leave room for them.
Architecture: 43 layers · 8 full attention (4 KV × 256) · 35 linear attention. Its KV cache is 5.4× smaller than if every layer used full attention, because only 8 of its 43 layers keep a KV cache; the others hold a small fixed-size state.
| Context per request | BF16 cache | FP8 cache |
|---|---|---|
| 8K tokens | 0.25 GiB | 0.13 GiB |
| 32K tokens | 1.00 GiB | 0.50 GiB |
| 128K tokens | 4.00 GiB | 2.00 GiB |
| 256K tokens | 8.00 GiB | 4.00 GiB |
Plus 71.6 MiB of recurrent state per request, whatever the context.
In BF16 the weights alone take 17.4 GiB (19 GB). In FP8 that is 10.5 GiB, and about 6.77 GiB with 4-bit quantization (Q4_K_M). Each request then adds KV cache: 0.25 GiB at 8K tokens and 1.00 GiB at 32K (BF16 cache).
Yes. In BF16 it fits on one RTX 4090 with an 8K-token context (17.7 GiB including the cache, with llama.cpp).
Yes. In BF16 (17.7 GiB with an 8K-token context) it fits on one RTX 5090, with llama.cpp.
Yes. In 4-bit (Q4_K_M with llama.cpp) it needs a Mac with 16 GB of unified memory or more, of which macOS gives the GPU about 11 GB, for one request with an 8K-token context. In 8-bit (Q8_0) it needs 24 GB or more.
Yes. In BF16 (17.7 GiB with an 8K-token context) it fits on a DGX Spark (128 GB) and on a Ryzen AI Max+ 395 with 128 GB, whose GPU gets up to 96 GB, with llama.cpp and one request.
On 2026-11-05, the cheapest on-demand setup is 1 × RTX A4000 (16 GB) with Q8_0 weights on Vast.ai, at $0.072 per hour (about $52.56 per month). Next: 1 × GeForce RTX 3090 (24 GB) with BF16 weights on Vast.ai, at $0.096 per hour. Sized for one 8K-token request; prices are checked every hour.
With one request and an 8K-token context: 1 in BF16, 1 in FP8 and 1 in 4-bit (vLLM, 90% of memory usable). Serving many users at once needs more memory for their KV caches: the calculator sizes that for you.
Yes, Qwen3.5 4B-Qwenputer supports tool (function) calling: it can ask your application to run a function and use the result.
Thinking mode you can switch on or off. With vLLM or SGLang, set it per request with "chat_template_kwargs": {"enable_thinking": false} (or true). Reasoning tokens count toward the context and the KV cache.
For a model card or a README: the 4-bit size and the smallest GPU it fits on, linked to this page. Data as JSON: /api/models/qwen3-5-4b-qwenputer.json.
[](https://studiotvai.com/vram-requirements/qwen3-5-4b-qwenputer)Same family (hybrid linear attention): Qwen3.5 4B · MiniCPM-V-4.7 1B · Raon-1 mini preview · Qwen3.5 9B · Qwen3.8 27B · Qwen3.6 35B-A3B · Clover 35B-A3B Preview · SenseNova Loop 9B Preview · Qwen3.5 122B-A10B · Qwen3.8-Flash-Next · LFM3 1.7B · LFM3-VL 2B · d1 2B · LFM2.5 8B-A1B · LFM2 24B-A2B · Ornith 1.5 35B-A3B · Ornith 1.5 9B · Clef · Clef Flash · d1 3B · JEV 27B VL · LightOnOCR-3 4B · Agnes-3.0-Qwen · Agnes-3.0-Flash · Qwen3.6 27B · Qwen3.5 27B · Qwen3.5 35B-A3B · MiniCPM-V-4.7 35B-A3B · LFM2.5 24B-A2B · Ornith-1.5 122B · Sage-A1.5 · MLExpert · Agents-A1 4B · Ling-3.0-flash-Sante · MiniCPM-OCR-1.0 · 6.8-Code2Video
Llama 3.1 8B · Llama 3 8B · Llama 3.2 3B · Nanbeige4.3 33B · Granite 4.2 30B · Granite 4.2 8B · Granite 4.2 3B · MiniCPM5 2B · MiniCPM5 1B · Ministral 3 14B · Ministral 3 8B · Ministral 3 3B · Devstral 2 123B · Devstral Small 2 24B · Mistral Medium 3.5 128B · Magistral Small 1.2 24B · Mistral Small 3 24B · Mistral Large 2411 123B · Pixtral Large 124B · Codestral 22B · Mistral Small 2409 22B · Pixtral 12B · Nemotron 3.5 Parse · Gemma 2 9B · Gemma 3 4B · LLaVA 1.6 Mistral 7B · Mistral Small 3.2 24B · Qwen3 32B · Llama 3.3 70B · Llama 3.1 405B · Hy-Memory 30B-A3B · Kimi VL A3B Instruct · Mixtral 8x7B · Unlimited-OCR · Unlimited-OCR 2 · Xing4.0 29B-A4B · Qwen3.5 4B · MiniCPM-V-4.7 1B · Raon-1 mini preview · Qwen3.5 9B · Qwen3.8 27B · Qwen3.6 35B-A3B · Clover 35B-A3B Preview · SenseNova Loop 9B Preview · Qwen3.5 122B-A10B · Qwen3.8-Flash-Next · LFM3 1.7B · LFM3-VL 2B · d1 2B · LFM2.5 8B-A1B · LFM2 24B-A2B · Gemma 4 12B · Gemma 4 31B · Gemma 4 26B-A4B · gpt-oss 20B · gpt-oss 120B · Llama 4 Scout 17B-16E · Llama 4 Maverick 17B-128E · Kolibri-1 · Recursive Flint · Nemotron Labs Dory 1.0 4B · Step 3.5 Flash · Step 3.7 Flash · Laguna S 2.1 · Laguna XS 2.1 · Command A+ · Inkling Small · Muse Glimmer 30B · Nemotron 3 Nano 30B-A3B · Nemotron 3 Super 120B-A12B · Nemotron 3 Ultra 550B-A55B · Mamba-Codestral 7B · Nemotron 3.5 Super VL 120B-A12B · Nemotron 3.5 Bolt 74B-A9B · MiniMax M3.1 Flash · MiniMax M3 · MiniMax M2.7 · DeepSeek R1 · Kimi K2.6 · GLM-5.3 · Atria Dawn · Atria Dawn Preview · Hy4 Preview · K-EXAONE 2.0 750B-A37B · Inkling · GLM-5.3-Flash · DeepSeek V4 Flash · DeepSeek V4 Pro · DeepSeek V4.1 Flash · MiMo-V2.6 Pro · MiMo-V2.6 Flash · Muse Spark Open · Beam 1.0 501B-A23B · Kimi K3 · Mistral Large 4 · Mistral Large 3 · Mistral Small 4 119B · Leanstral 1.5 119B-A6B · Mixtral 8x22B · Qwen3-Coder-Next 80B-A3B · Qwen3-Coder 30B-A3B · Qwen3-Coder 480B-A35B · DeepSeek V4 Flash Vision Exp · Qwen3.8 2.4T-A95B · Ornith 1.5 397B · Ornith 1.5 35B-A3B · Ornith 1.5 9B · Intern S2 397B · Clef · Clef Flash · d1 3B · Humanizer · JEV 27B VL · Mellum2.1 12B-A2.5B · LightOnOCR-3 4B · Spark-X2.5 4B · Agnes-3.0-Qwen · Youtu-Parsing-Omni · Agnes-3.0-Flash · Mistral 7B v0.3 · Mistral Small 3.1 24B · Mistral Nemo · Ministral 8B · DeepSeek V3 · DeepSeek R1 Qwen3 8B · DeepSeek R1 Distill Qwen 32B · DeepSeek V3.1 · DeepSeek R1 Distill Qwen 14B · DeepSeek R1 Distill Qwen 7B · DeepSeek R1 Distill Llama 8B · Qwen3 4B · Qwen3 8B · Qwen2.5 7B · Qwen3 14B · Qwen2.5 3B · Qwen3.6 27B · Qwen3.5 27B · Qwen2.5-Coder 7B · Qwen2.5 32B · Qwen2.5-Coder 14B · Qwen3.5 35B-A3B · Qwen3 30B-A3B · Qwen2.5 14B · Qwen2.5-Coder 32B · Kimi K2.7 Code · Kimi K2.5 · Kimi K2 · GLM-5.2 · GLM-5.1 · GLM-5 · MiniMax M2.5 · MiniMax M2 · MiMo-V2.5 · Ornith-1.0 397B · Qwen2.5-Coder 3B · Qwen3 235B-A22B · Qwen3.5 397B-A17B · Gemma 4 E4B · Gemma 4 E2B · Nemotron 3 Nano 4B · Nemotron 3.5 Lightning 30B-A3B · MiniCPM-V-4.7 35B-A3B · LFM2.5 24B-A2B · GLM-4.7 · Qwen3-VL 8B · Qwen2.5-VL 7B · Qwen3-VL 4B · Qwen2.5-VL 3B · Qwen2.5-VL 32B · Qwen2-VL 7B · GLM-4.5-Air · GLM-4.5 · phi-4 · Phi-3.5-mini · Phi-4-mini · Phi-3-mini-128k · Ling-3.1-flash · Step 5 Preview · Hy3 · Ornith-1.5 122B · Spark-X2.6 4B · Sage-A1.5 · MLExpert · Agents-A1 4B · Ling-3.0-flash-Sante · MiniCPM-OCR-1.0 · WeVisOne-VL 4B · Granite 4.1 3B · Qwen3-VL 30B-A3B · Mage-VL-Cowork · Spark-X2.6 1.7B · Kimi K3.1 Pro · 6.8-Code2Video
Estimates, not guarantees: computed from the official config.json with the same engine as the LLM VRAM Calculator. Real usage depends on your engine version and settings.