FitMyLLM
Alibaba/Dense

AlibabaQwen 3.5 4B

Qwen 3.5 4B — multimodal agent model. Beats models 2-3x its size on reasoning and coding.

chatcodingreasoningmultilingualvisionmath
4.66B
Parameters
256K
Context length
18
Benchmarks
6
Quantizations
2.0M
HF downloads
Architecture
Dense
Released
2026-03-01
Layers
32
KV Heads
4
Head Dim
256
Family
qwen

Quantization Options

Context length:
QuantBitsVRAM @ 16KQuality
Q4_K_M4.89
4.8 GB
3.3 + 1.5 KV
good
Q5_K_S5.57
5.2 GB
3.7 + 1.5 KV
good
Q5_K_M5.7
5.3 GB
3.8 + 1.5 KV
good
Q6_K6.56
5.8 GB
4.3 + 1.5 KV
excellent
Q8_08.5
6.9 GB
5.4 + 1.5 KV
lossless
FP1616
11.3 GB
9.8 + 1.5 KV
lossless

Select your GPU above to see speed estimates and compatibility for each quantization.

Deploying for a team or in production? Size GPUs, cost & scaling in Enterprise →
READY TO RUN THIS?RENT BY THE HOUR

RENT A GPU AND RUN QWEN 3.5 4B NOW

Spin up an A100 / H100 / 4090 in ~60s. Pay by the second. Cancel anytime.

Community Ratings

Loading ratings...

Benchmarks (18)

τ²-Bench92.1
IFEval89.8
MMBench89.4
MMLU-PRO79.1
MMMU77.6
GPQA Diamond76.2
LiveCodeBench55.8
AA Long Context55.7
IFBench52.0
MATH49.6
BBH34.9
AA Intelligence27.1
Terminal-Bench18.2
AA Coding17.5
SciCode16.1
MUSR8.7
HLE7.8
GPQA6.4

Run this model

Easiest way to get started·Beginners
DOCS ↗
curl -fsSL https://ollama.com/install.sh | sh
$ollama run qwen3.5:4b-q4_K_M

Downloads and runs automatically. Add --verbose for speed stats.

▸ SETUP GUIDE
>_

Auto-setup with fitmyllm CLI

Detects your GPU, recommends the best model, downloads it, and starts chatting — zero config. Benchmarks your speed and contributes anonymous data to improve predictions.

pip install fitmyllmthen run fitmyllmLearn more
Auto-detect GPULive tok/s in chatSpeed benchmarks9 inference engines

GPUs that can run this model

At Q4_K_M quantization. Sorted by minimum VRAM.

NVIDIA Tesla C1080
4 GB VRAM • 102 GB/s
NVIDIA
NVIDIA Tesla K10
4 GB VRAM • 160 GB/s
NVIDIA
NVIDIA Tesla M4
4 GB VRAM • 88 GB/s
NVIDIA
AMD Radeon Instinct MI8
4 GB VRAM • 512 GB/s
AMD
NVIDIA RTX A1000 Embedded
4 GB VRAM • 224 GB/s
NVIDIA
NVIDIA RTX A2000 Embedded
4 GB VRAM • 192 GB/s
NVIDIA
Intel Arc A310
4 GB VRAM • 124 GB/s
INTEL
$79
Intel Arc A350
4 GB VRAM • 124 GB/s
INTEL
$99

Find the best GPU for Qwen 3.5 4B

Build Hardware for Qwen 3.5 4B
▸ SPEC SHEET

Qwen 3.5 4B4.66B Dense.

▸ SPECIFICATIONS
PARAMETERS
4.66B
ARCHITECTURE
Dense Transformer
CONTEXT LENGTH
256K tokens
CAPABILITIES
chat, coding, reasoning, multilingual, vision, math
RELEASE DATE
2026-03-01
PROVIDER
Alibaba
FAMILY
qwen
▸ VRAM REQUIREMENTS
QUANTBPWVRAMQUALITY
Q4_K_M4.893.3 GB94%
Q5_K_S5.573.7 GB96%
Q5_K_M5.73.8 GB96%
Q6_K6.564.3 GB97%
Q8_08.55.4 GB100%
FP16169.8 GB100%
§ 01BENCHMARK SCORES
MMLU-PRO79.1
MATH49.6
IFEval89.8
BBH34.9
MMMU77.6
GPQA6.4
MUSR8.7
MMBench89.4
GPQA Diamond76.2
HLE7.8
AA Intelligence27.1
AA Coding17.5
LiveCodeBench55.8
aa_ifbench52.0
aa_terminal_bench18.2
aa_tau292.1
aa_scicode16.1
aa_lcr55.7
§ 02RUN COMMAND

Run Qwen 3.5 4B locally with Ollama — needs 3.3 GB VRAM at Q4_K_M:

$ollama run qwen3.5:4b
§ 03COMPATIBLE GPUs
30 @ Q4_K_M