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 64K 128K
Quant Bits VRAM @ 16K Quality Q3_K_M 4 7.5 GB
6.0 + 1.5 KV
low Q3_K_L 4.3 7.9 GB
6.4 + 1.5 KV
moderate IQ4_XS 4.46 8.1 GB
6.6 + 1.5 KV
moderate Q4_K_S 4.67 8.4 GB
6.9 + 1.5 KV
moderate Q4_K_M 4.89 8.7 GB
7.2 + 1.5 KV
good Q5_K_S 5.57 9.6 GB
8.1 + 1.5 KV
good Q5_K_M 5.7 9.8 GB
8.3 + 1.5 KV
good Q6_K 6.56 11.0 GB
9.5 + 1.5 KV
excellent Q8_0 8.5 13.7 GB
12.2 + 1.5 KV
lossless FP16 16 24.0 GB
22.5 + 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 LLAMA-3.2-11B-VISION-INSTRUCT 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 Q3_K_M — 6.0 GB VRAM Q3_K_L — 6.4 GB VRAM IQ4_XS — 6.6 GB VRAM Q4_K_S — 6.9 GB VRAM Q4_K_M — 7.2 GB VRAM Q5_K_S — 8.1 GB VRAM Q5_K_M — 8.3 GB VRAM Q6_K — 9.5 GB VRAM Q8_0 — 12.2 GB VRAM FP16 — 22.5 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 llama3.2-vision:11b-instruct-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 Llama-3.2-11B-Vision-Instruct
Build Hardware for Llama-3.2-11B-Vision-Instruct Llama 3.2 11B Vision Instruct — instruction-tuned for visual Q&A 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
Llama-3.2-11B-Vision-Instruct — 11B Dense. ▸ SPECIFICATIONS
PARAMETERS 11B
ARCHITECTURE Dense Transformer
CONTEXT LENGTH 128K tokens
CAPABILITIES vision, chat
RELEASE DATE 2024-09-25
PROVIDER Meta
FAMILY llama ▸ VRAM REQUIREMENTS
QUANT BPW VRAM QUALITY Q3_K_M 4 6.0 GB 88% Q3_K_L 4.3 6.4 GB 90% IQ4_XS 4.46 6.6 GB 92% Q4_K_S 4.67 6.9 GB 93% Q4_K_M 4.89 7.2 GB 94% Q5_K_S 5.57 8.1 GB 96% Q5_K_M 5.7 8.3 GB 96% Q6_K 6.56 9.5 GB 97% Q8_0 8.5 12.2 GB 100% FP16 16 22.5 GB 100%
§ 01 BENCHMARK SCORES
HumanEval 45.0
MMLU-PRO 55.0
MATH 35.0
IFEval 70.0
BBH 55.0
MMMU 50.7
GPQA 35.0
MUSR 18.0
MMBench 76.8
LiveCodeBench 11.0
AIME 1.7
MATH-500 1.7
GPQA Diamond 22.1
HLE 5.2
§ 02 RUN COMMAND
Run Llama-3.2-11B-Vision-Instruct locally with Ollama — needs 7.2 GB VRAM at Q4_K_M:
$ ollama run llama3.2-vision:11b
§ 03 COMPATIBLE GPUs
30 @ Q4_K_M Feedback 