GPU VRAM calculator

Run Qwen3-VL-8B-Instruct locally: hardware requirements

Qwen3-VL-8B-Instruct needs about 24.1 GB of GPU memory for one person at 32K context, and about 9.0 GB more for each extra person at 64K. For 80 users working as AI agents, plan on 164 GB, which fits one B200.

Alibaba (Qwen), China. From its files on Hugging Face, checked 2026-11-05.

Context per person: more context lets the model work with longer conversations and more files, and needs more GPU memory for every person. As a guide: 8K for short chat, 32K for chat over documents, 64K for AI agents, whose instructions, tools and files all count.

80 users means about 16 people working at the same moment. That share is a planning assumption; if you know yours, choose the total that matches it.

Work it out for your team: Qwen/Qwen3-VL-8B-Instruct

On Hugging Face

Alibaba (Qwen), China · 8.8B parameters · License: apache-2.0

What you need for 16 active users

1 × B200

It holds the model and up to 16 active users at 64K context.

The server

384 GB RAM
32 CPU cores
250 GB NVMe

Rules of thumb

GPU memory in total
164 GB
Weights
17.5 GB (published files)
Conversation cache, all active
144 GB
Engine (vLLM)
3.0 GB

Your options by card

80 users means about 16 using the model at any moment. The others are not waiting; they are simply doing something else. If more become active than a build handles, they queue. 1 card too small for this model is left out.

CardMemoryActive users, one serverFor your 16 active
B200180 GB161 × B200
A10080 GB233 × A100
GeForce RTX 309024 GB28 servers, each with 2 × GeForce RTX 3090
GeForce RTX 409024 GB28 servers, each with 2 × GeForce RTX 4090
GeForce RTX 508016 GB116 servers, each with 2 × GeForce RTX 5080
GeForce RTX 509032 GB44 servers, each with 2 × GeForce RTX 5090
H10080 GB233 × H100
H200 NVL141 GB272 × H200 NVL
L40S48 GB184 × L40S
RTX 4000 Ada20 GB63 servers, each with 4 × RTX 4000 Ada
RTX 6000 Ada48 GB184 × RTX 6000 Ada
RTX A500024 GB73 servers, each with 4 × RTX A5000
RTX PRO 4000 Blackwell24 GB73 servers, each with 4 × RTX PRO 4000 Blackwell
RTX PRO 4500 Blackwell32 GB112 servers, each with 4 × RTX PRO 4500 Blackwell
RTX PRO 5000 Blackwell48 GB184 × RTX PRO 5000 Blackwell
RTX PRO 6000 Blackwell96 GB182 × RTX PRO 6000 Blackwell

These are memory limits. A card can hold more people than it can answer quickly, so test your own workload before you buy. More than one card in a server needs an engine that splits the model.

Qwen3-VL-8B-Instruct by team size

AI agents, 1 user in 5 active at a time, 64K context, as published.

Total usersAt onceGPU memoryWhat you need
20455.6 GB1 × A100
8016164 GB1 × B200
12024236 GB2 × H200 NVL
500100920 GB6 × B200
  • Qwen3-8BAlibaba (Qwen) · 8.2B · 109 GB for 16 at once
  • Qwen3.5-9BAlibaba (Qwen) · 9.7B · 53.3 GB for 16 at once
  • Qwen3-14BAlibaba (Qwen) · 15B · 131 GB for 16 at once
  • Qwen3.5-4BAlibaba (Qwen) · 4.7B · 43.8 GB for 16 at once
  • Qwen3-4BAlibaba (Qwen) · 4.0B · 101 GB for 16 at once
  • Qwen3.8-27BAlibaba (Qwen) · 28B · 120 GB for 16 at once

All models

How this is worked out

GPU memory = weights + a conversation cache for each person + the serving engine, from this model's files on Hugging Face. These are memory limits, not speed. The full method and its limits.

Questions

How much GPU memory does Qwen3-VL-8B-Instruct need?

Qwen3-VL-8B-Instruct needs about 24.1 GB of GPU memory for one person at 32K context, and about 9.0 GB more for each extra person at 64K. For 80 users working as AI agents, plan on 164 GB, which fits one B200.

How many active users can Qwen3-VL-8B-Instruct handle on one 96 GB GPU?

By memory, about 7 active users at once at 64K context on a card such as the RTX PRO 6000 Blackwell. Speed can set a lower limit, so load-test before you buy.

Can Qwen3-VL-8B-Instruct run on a 24 GB GPU such as the RTX 4090?

Not as published for one person at 32K context: it needs about 24.1 GB. A 4-bit or 8-bit build may fit, if the model has one.

What license does Qwen3-VL-8B-Instruct use?

Hugging Face lists its license as apache-2.0. Check the license on the model's page before you use it.

Running a model on your own servers?

NeueCode is private autonomous intelligence for enterprise: it takes on your organization's tasks on your own GPUs, with nothing leaving your network, a person approving every action, and all of it on record. It runs each model on one card, so read the single-card figures above.

What you need to run NeueCode