FITMYLLM · JULY 28, 2026 · OPEN METHODOLOGY · COMMUNITY BENCHMARKS
LIVE · UPDATED DAILY
358B
Parameters (32B active)
Quantization Options Select your GPU for speed estimates Context length: 4K 8K 16K 32K 64K 128K 193K
Quant Bits VRAM @ 16K Quality IQ2_XXS 2.38 111.3 GB
107.0 + 4.3 KV
low IQ2_M 2.93 135.9 GB
131.6 + 4.3 KV
low Q2_K 3.16 146.2 GB
141.9 + 4.3 KV
low IQ3_XXS 3.25 150.2 GB
145.9 + 4.3 KV
low IQ3_XS 3.5 161.4 GB
157.1 + 4.3 KV
low Q3_K_S 3.64 167.7 GB
163.4 + 4.3 KV
low IQ3_M 3.76 173.1 GB
168.7 + 4.3 KV
low Q3_K_M 4 183.8 GB
179.5 + 4.3 KV
low Q3_K_L 4.3 197.2 GB
192.9 + 4.3 KV
moderate IQ4_XS 4.46 204.4 GB
200.1 + 4.3 KV
moderate Q4_K_S 4.67 213.8 GB
209.5 + 4.3 KV
moderate Q4_K_M 4.89 223.6 GB
219.3 + 4.3 KV
good Q5_K_S 5.57 254.1 GB
249.7 + 4.3 KV
good Q5_K_M 5.7 259.9 GB
255.6 + 4.3 KV
good Q6_K 6.56 298.4 GB
294.0 + 4.3 KV
excellent Q8_0 8.5 385.2 GB
380.9 + 4.3 KV
lossless FP16 16 720.8 GB
716.5 + 4.3 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 ~111 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 4.7 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 — 107.0 GB VRAM IQ2_M — 131.6 GB VRAM Q2_K — 141.9 GB VRAM IQ3_XXS — 145.9 GB VRAM IQ3_XS — 157.1 GB VRAM Q3_K_S — 163.4 GB VRAM IQ3_M — 168.7 GB VRAM Q3_K_M — 179.5 GB VRAM Q3_K_L — 192.9 GB VRAM IQ4_XS — 200.1 GB VRAM Q4_K_S — 209.5 GB VRAM Q4_K_M — 219.3 GB VRAM Q5_K_S — 249.7 GB VRAM Q5_K_M — 255.6 GB VRAM Q6_K — 294.0 GB VRAM Q8_0 — 380.9 GB VRAM FP16 — 716.5 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:358b-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
GPUs that can run this model At Q4_K_M quantization. Sorted by minimum VRAM.
Find the best GPU for GLM 4.7
Build Hardware for GLM 4.7 GLM 4.7 — latest in the series with stronger coding and agentic capabilities.
Read full model card ▸ COLOPHON FITMYLLM · INDEPENDENT · DATA-DRIVEN
FITMYLLM · EST. 2025 · © 2026
RECOMMENDATIONS FROM PUBLISHED MATH, CORRECTED BY THE COMMUNITY — 30.
▸ SPEC SHEET
GLM 4.7 — 358B MoE. ▸ SPECIFICATIONS
PARAMETERS 358B (32B active)
ARCHITECTURE Mixture of Experts
CONTEXT LENGTH 193K tokens
CAPABILITIES chat, coding, reasoning, tool_use
RELEASE DATE 2026-01-29
PROVIDER Zhipu AI
FAMILY glm ▸ VRAM REQUIREMENTS
QUANT BPW VRAM QUALITY IQ2_XXS 2.38 107.0 GB 65% IQ2_M 2.93 131.6 GB 75% Q2_K 3.16 141.9 GB 78% IQ3_XXS 3.25 145.9 GB 82% IQ3_XS 3.5 157.1 GB 84% Q3_K_S 3.64 163.4 GB 85% IQ3_M 3.76 168.7 GB 86% Q3_K_M 4 179.5 GB 88% Q3_K_L 4.3 192.9 GB 90% IQ4_XS 4.46 200.1 GB 92% Q4_K_S 4.67 209.5 GB 93% Q4_K_M 4.89 219.3 GB 94% Q5_K_S 5.57 249.7 GB 96% Q5_K_M 5.7 255.6 GB 96% Q6_K 6.56 294.0 GB 97% Q8_0 8.5 380.9 GB 100% FP16 16 716.5 GB 100%
§ 01 BENCHMARK SCORES
MMLU-PRO 84.3
GPQA Diamond 85.7
LiveCodeBench 84.9
AIME 95.7
MATH-500 48.0
HLE 24.8
AA Intelligence 42.1
AA Coding 36.3
AA Math 95.0
aa_ifbench 67.9
aa_terminal_bench 31.8
aa_tau2 95.9
aa_scicode 45.1
aa_lcr 64.0
§ 03 COMPATIBLE GPUs
5 @ Q4_K_M Feedback 