FITMYLLM · AUGUST 17, 2026 · OPEN METHODOLOGY · COMMUNITY BENCHMARKS
LIVE · UPDATED DAILY
744B
Parameters (40B active)
Quantization Options Select your GPU for speed estimates Context length: 4K 8K 16K 32K 64K 128K 198K
Quant Bits VRAM @ 16K Quality IQ2_XXS 2.38 236.5 GB
221.8 + 14.6 KV
low IQ2_M 2.93 287.6 GB
273.0 + 14.6 KV
low Q2_K 3.16 309.0 GB
294.4 + 14.6 KV
low IQ3_XXS 3.25 317.4 GB
302.7 + 14.6 KV
low IQ3_XS 3.5 340.6 GB
326.0 + 14.6 KV
low Q3_K_S 3.64 353.6 GB
339.0 + 14.6 KV
low IQ3_M 3.76 364.8 GB
350.2 + 14.6 KV
low Q3_K_M 4 387.1 GB
372.5 + 14.6 KV
low Q3_K_L 4.3 415.0 GB
400.4 + 14.6 KV
moderate IQ4_XS 4.46 429.9 GB
415.3 + 14.6 KV
moderate Q4_K_S 4.67 449.4 GB
434.8 + 14.6 KV
moderate Q4_K_M 4.89 469.9 GB
455.3 + 14.6 KV
good Q5_K_S 5.57 533.1 GB
518.5 + 14.6 KV
good Q5_K_M 5.7 545.2 GB
530.6 + 14.6 KV
good Q6_K 6.56 625.2 GB
610.6 + 14.6 KV
excellent Q8_0 8.5 805.6 GB
791.0 + 14.6 KV
lossless FP16 16 1503.1 GB
1488.5 + 14.6 KV
lossless
Select your GPU above to see speed estimates and compatibility for each quantization.
Too big for a single GPU — plan a multi-GPU deployment
Even the lightest quant needs ~236 GB. Size GPUs, replicas, TCO and scaling for a production setup. Open in Enterprise →
▸ READY TO RUN THIS? RENT BY THE HOUR
RENT A GPU AND RUN GLM-5 744B 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 IQ2_XXS — 221.8 GB VRAM IQ2_M — 273.0 GB VRAM Q2_K — 294.4 GB VRAM IQ3_XXS — 302.7 GB VRAM IQ3_XS — 326.0 GB VRAM Q3_K_S — 339.0 GB VRAM IQ3_M — 350.2 GB VRAM Q3_K_M — 372.5 GB VRAM Q3_K_L — 400.4 GB VRAM IQ4_XS — 415.3 GB VRAM Q4_K_S — 434.8 GB VRAM Q4_K_M — 455.3 GB VRAM Q5_K_S — 518.5 GB VRAM Q5_K_M — 530.6 GB VRAM Q6_K — 610.6 GB VRAM Q8_0 — 791.0 GB VRAM FP16 — 1488.5 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 glm:744b-q4_K_MCOPY
Tag may need adjustment — check ollama.com/library/glm for available tags.
▸ 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
Find the best GPU for GLM-5 744B
Build Hardware for GLM-5 744B ▸ COLOPHON FITMYLLM · INDEPENDENT · DATA-DRIVEN
FITMYLLM · EST. 2025 · © 2026
RECOMMENDATIONS FROM PUBLISHED MATH, CORRECTED BY THE COMMUNITY — 30.
▸ SPEC SHEET
GLM-5 744B — 744B MoE. ▸ SPECIFICATIONS
PARAMETERS 744B (40B active)
ARCHITECTURE Mixture of Experts
CONTEXT LENGTH 198K tokens
CAPABILITIES chat, coding, reasoning, multilingual
RELEASE DATE 2026-02-01
PROVIDER Zhipu AI
FAMILY glm ▸ VRAM REQUIREMENTS
QUANT BPW VRAM QUALITY IQ2_XXS 2.38 221.8 GB 65% IQ2_M 2.93 273.0 GB 75% Q2_K 3.16 294.4 GB 78% IQ3_XXS 3.25 302.7 GB 82% IQ3_XS 3.5 326.0 GB 84% Q3_K_S 3.64 339.0 GB 85% IQ3_M 3.76 350.2 GB 86% Q3_K_M 4 372.5 GB 88% Q3_K_L 4.3 400.4 GB 90% IQ4_XS 4.46 415.3 GB 92% Q4_K_S 4.67 434.8 GB 93% Q4_K_M 4.89 455.3 GB 94% Q5_K_S 5.57 518.5 GB 96% Q5_K_M 5.7 530.6 GB 96% Q6_K 6.56 610.6 GB 97% Q8_0 8.5 791.0 GB 100% FP16 16 1488.5 GB 100%
§ 01 BENCHMARK SCORES
SWE-bench 77.8
AIME 92.7
GPQA Diamond 86.0
§ 02 RUN COMMAND
Run GLM-5 744B locally with Ollama — needs 455.3 GB VRAM at Q4_K_M:
$ ollama run glm-5
Feedback 