FITMYLLM · JULY 28, 2026 · OPEN METHODOLOGY · COMMUNITY BENCHMARKS
LIVE · UPDATED DAILY
Quantization Options Select your GPU for speed estimates Context length: 4K 8K 16K 32K
Quant Bits VRAM @ 16K Quality IQ3_XXS 3.25 16.5 GB
13.5 + 3.0 KV
low IQ3_XS 3.5 17.5 GB
14.5 + 3.0 KV
low Q3_K_S 3.64 18.0 GB
15.0 + 3.0 KV
low IQ3_M 3.76 18.5 GB
15.5 + 3.0 KV
low Q3_K_M 4 19.5 GB
16.5 + 3.0 KV
low Q3_K_L 4.3 20.7 GB
17.7 + 3.0 KV
moderate IQ4_XS 4.46 21.3 GB
18.3 + 3.0 KV
moderate Q4_K_S 4.67 22.2 GB
19.2 + 3.0 KV
moderate Q4_K_M 4.89 23.0 GB
20.0 + 3.0 KV
good Q5_K_S 5.57 25.8 GB
22.8 + 3.0 KV
good Q5_K_M 5.7 26.3 GB
23.3 + 3.0 KV
good Q6_K 6.56 29.7 GB
26.7 + 3.0 KV
excellent Q8_0 8.5 37.5 GB
34.5 + 3.0 KV
lossless FP16 16 67.5 GB
64.5 + 3.0 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 1.5 32B NOW
Spin up an A100 / H100 / 4090 in ~60s. Pay by the second. Cancel anytime.
Community Ratings Chat Coding Reasoning Creative Vision Roleplay Agentic
Loading ratings...
Run this model IQ3_XXS — 13.5 GB VRAM IQ3_XS — 14.5 GB VRAM Q3_K_S — 15.0 GB VRAM IQ3_M — 15.5 GB VRAM Q3_K_M — 16.5 GB VRAM Q3_K_L — 17.7 GB VRAM IQ4_XS — 18.3 GB VRAM Q4_K_S — 19.2 GB VRAM Q4_K_M — 20.0 GB VRAM Q5_K_S — 22.8 GB VRAM Q5_K_M — 23.3 GB VRAM Q6_K — 26.7 GB VRAM Q8_0 — 34.5 GB VRAM FP16 — 64.5 GB VRAM
Ollama llama.cpp vLLM LM Studio KoboldCpp Jan Docker
▸ Easiest way to get started · Beginners
DOCS ↗ curl -fsSL https://ollama.com/install.sh | shCOPY
$ ollama run qwen:32b-chat-v1.5-q4_K_MCOPY
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.
Auto-detect GPU Live tok/s in chat Speed benchmarks 9 inference engines
GPUs that can run this model At Q4_K_M quantization. Sorted by minimum VRAM.
Find the best GPU for Qwen 1.5 32B
Build Hardware for Qwen 1.5 32B Qwen 1.5 32B — large model for demanding Chinese-English tasks.
Read full model card ▸ COLOPHON FITMYLLM · INDEPENDENT · DATA-DRIVEN
FITMYLLM · EST. 2025 · © 2026
RECOMMENDATIONS FROM PUBLISHED MATH, CORRECTED BY THE COMMUNITY — 30.
▸ SPEC SHEET
Qwen 1.5 32B — 32B Dense. ▸ SPECIFICATIONS
PARAMETERS 32B
ARCHITECTURE Dense Transformer
CONTEXT LENGTH 32K tokens
CAPABILITIES chat
RELEASE DATE 2024-02-04
PROVIDER Alibaba
FAMILY qwen ▸ VRAM REQUIREMENTS
QUANT BPW VRAM QUALITY IQ3_XXS 3.25 13.5 GB 82% IQ3_XS 3.5 14.5 GB 84% Q3_K_S 3.64 15.0 GB 85% IQ3_M 3.76 15.5 GB 86% Q3_K_M 4 16.5 GB 88% Q3_K_L 4.3 17.7 GB 90% IQ4_XS 4.46 18.3 GB 92% Q4_K_S 4.67 19.2 GB 93% Q4_K_M 4.89 20.0 GB 94% Q5_K_S 5.57 22.8 GB 96% Q5_K_M 5.7 23.3 GB 96% Q6_K 6.56 26.7 GB 97% Q8_0 8.5 34.5 GB 100% FP16 16 64.5 GB 100%
§ 01 BENCHMARK SCORES
HumanEval 87.2
MMLU-PRO 51.9
MATH 62.5
IFEval 83.5
BBH 56.5
GPQA 11.7
MUSR 13.5
MBPP 77.0
BigCodeBench 45.0
§ 02 RUN COMMAND
Run Qwen 1.5 32B locally with Ollama — needs 20.0 GB VRAM at Q4_K_M:
$ ollama run qwen:32b
§ 03 COMPATIBLE GPUs
30 @ Q4_K_M Feedback 