Qwen
ChatCodingAI codingReasoningMulti-purpose8.2B
Chat

Qwen3-8B Instruct

Qwen · Apache 2.0

Qwen 3 dense 8B instruct from Alibaba (~8.2B parameters, ~6.95B non-embedding). Capable chat, coding, and reasoning; 32K default context with 128K max. Apache 2.0.

Parameters
8.2B
Architecture
Dense
Context
131,072 tokens
Released
2025-04-29
Engines
llama.cpp, ollama, vLLM
Builder Tools
Continue, LM Studio, Open WebUI

Parameters

8.2B

VRAM

6.5 GB

Context

128K

Formats

4

GPUs

20

Qwen3-8B Instruct (8.2B) requires 6.5 GB VRAM at recommended quality (Q5_K_M). At efficient quality (Q4_K_M), it fits in 5.5 GB VRAM, making it compatible with the NVIDIA RTX 4060 Laptop (40-60W). On NVIDIA Grace Blackwell Ultra GB300, expect approximately 340 tok/s at Q8_0. For the best experience, Starter AI Desktop ($582) is recommended.

Source: OwnRig methodology

VRAM (Recommended)

6.5 GB

Quantization

Q5_K_M

File Size

5.9 GB

Max Context

128K tokens

Primary Use

Chat

Memory

VRAM Requirements

QualityQuantizationVRAMFile Size
fullQ8_09.5 GB8.7 GB
recommendedQ5_K_M6.5 GB5.9 GB
efficientQ4_K_M5.5 GB5 GB
compressedQ3_K_M4.5 GB4.1 GB
Scaling

Context Length Impact

KV cache VRAM at Q5_K_M quality. Longer context = more memory.

ContextKV CacheTotal VRAM
2K102 MB6.6 GB
4K205 MB6.7 GB
8K512 MB7 GB
16K1 GB7.5 GB
32K1.9 GB8.4 GB
64K3.8 GB10.3 GB
128K7.7 GB14.2 GB

Compatible GPUs

20 devices
NVIDIA Grace Blackwell Ultra GB300Q8_0340 tok/sExcellent
NVIDIA GeForce RTX 4090Q8_083 tok/sExcellent
AMD Radeon Pro W7900Q8_090 tok/sExcellent
NVIDIA RTX PRO 6000 BlackwellQ8_0120 tok/sExcellent
NVIDIA RTX PRO 6000 Blackwell Max-QQ8_0110 tok/sExcellent
NVIDIA GeForce RTX 3080 10GBQ8_032 tok/sGood
NVIDIA GeForce RTX 4070 SuperQ8_032 tok/sGood
NVIDIA GeForce RTX 4070 Ti 12GBQ8_029 tok/sGood
AMD Radeon RX 7900 XTXQ8_071 tok/sGood
NVIDIA GeForce RTX 3060 12GBQ8_020 tok/sAcceptable
NVIDIA GeForce RTX 4060 8GBQ5_K_M24 tok/sAcceptable
NVIDIA RTX 4060 Laptop (40-60W)Q5_K_M14 tok/sAcceptable
NVIDIA RTX 4070 Laptop (80-115W)Q5_K_M16 tok/sAcceptable
NVIDIA RTX 4080 Laptop (120-150W)Q8_021 tok/sAcceptable
AMD Radeon RX 7600Q5_K_M19 tok/sAcceptable
AMD Radeon RX 9070Q8_0Acceptable
AMD Radeon RX 9060 XT 16GBQ8_0Acceptable
NVIDIA GeForce RTX 5060 8GBQ5_K_M28 tok/sAcceptable
Apple M3 Pro (18GB Unified)Q8_08 tok/sMarginal
AMD Radeon RX 9060 XT 8GBQ8_0Not viable

Showing 20 of 20 entries

Builder Context

Qwen3-8B Instruct is commonly used with Continue, LM Studio, Open WebUI. For an AI coding workflow, pair it with an embedding model like nomic-embed-text for local RAG.

FAQ

Frequently Asked Questions

How much VRAM does Qwen3-8B Instruct need?
Qwen3-8B Instruct requires 6.5 GB VRAM at recommended quality (Q5_K_M). At lower quality settings, it can fit in as little as 4.5 GB.
What is the best GPU for Qwen3-8B Instruct?
The NVIDIA Grace Blackwell Ultra GB300 delivers the best performance for Qwen3-8B Instruct, achieving 340 tok/s at Q8_0 with an excellent rating.
What quantization should I use for Qwen3-8B Instruct?
For the best quality, use Q5_K_M (6.5 GB VRAM). If your GPU has limited VRAM, Q3_K_M (4.5 GB) is the most efficient option with acceptable quality.
Is Qwen3-8B Instruct good for coding?
Yes. Qwen3-8B Instruct is used with Continue, LM Studio, Open WebUI for local AI coding. For the best coding experience, pair it with an embedding model for local RAG.
All models

Data confidence: community. Source

VRAM requirements are calculated from model parameters and may vary by inference engine, context length, and batch size. Performance estimates are based on community benchmarks and should be verified for your specific configuration.Qwen is a trademark of its respective owner. OwnRig is not affiliated with or endorsed by the model creator.