GLM-4
Jun 2024
GLM-4Compare inon
Benchmarks of GLM-411
BenchmarkByScoreRank
No benchmark matches
—per task?
price not verifiedprice
—output speed?
—context
Ranking?
Alternatives?
ModelScoreCostPer taskWhat changes
BetterGPT-5 Minihighwins 69%—$0.053
OpenNVIDIA Nemotron 3.5 Lightning 30B A3B NVFP4wins 67%——Memory (4-bit): 4.7 GB → 11 GBBetterGPT-6 Lunamaxwins 71%—$0.068Function calling only with none · Chat Completions+3
BetterClaude Haiku 5.5highwins 72%—$0.079Remove sampling parameters+5BetterGPT-5.6 Lunaxhighwins 76%—$0.085ModelScoreCostPer taskWhat changes
Nothing here yet.
ModelScoreCostPer taskWhat changes
GPT-5 Minihighwins 69%—$0.053GPT-6 Lunamaxwins 71%—$0.068Function calling only with none · Chat Completions+3Claude Haiku 5.5highwins 72%—$0.079Remove sampling parameters+5GPT-5.6 Lunaxhighwins 76%—$0.085GPT-5.6 Terraxhighwins 77%—$0.63Prompts with >272K input tokensModelScoreCostPer taskWhat changes
NVIDIA Nemotron 3.5 Lightning 30B A3B NVFP4Openwins 67%——Memory (4-bit): 4.7 GB → 11 GB
Gemma 3 12BunstatedOpenwins 65%—$0.05 / $0.10Memory (4-bit): 4.7 GB → 6.1 GB
Granite 4.2 8BunstatedOpenwins 63%—$0.06 / $0.25Granite 4.2 3BunstatedOpenwins 61%—$0.03 / $0.12Gemma 3n E4B InstructunstatedOpenwins 61%——ModelScoreCostPer taskWhat changes
GLM 5.3maxOpenwins 80%—$2.01Memory (4-bit): 4.7 GB → 452 GB+4GLM 5.1unstatedOpenwins 79%—$1.40 / $4.40Memory (4-bit): 4.7 GB → 452 GBGLM 5.2maxOpenwins 79%—$1.28Memory (4-bit): 4.7 GB → 452 GBGLM 5.3 FlashunstatedOpenwins 79%—$0.15 / $0.50Memory (4-bit): 4.7 GB → 193 GB+5GLM 5unstatedOpenwins 78%—$1 / $3.20Memory (4-bit): 4.7 GB → 452 GBGLM 5V Turbounstatedwins 77%—$1.20 / $4GLM 4.7unstatedOpenwins 77%—$0.60 / $2.20Memory (4-bit): 4.7 GB → 215 GBGLM 4.5unstatedOpenwins 75%—$0.60 / $2.20Memory (4-bit): 4.7 GB → 215 GBAll Z.ai models →Run it yourself
estimateFits on✓ 8 GB (phone, laptop) · compressed✓ 16 GB · 8-bit✓ 24 GB GPU · full quality✓ 48 GB GPU · full quality✓ 80 GB GPU · full quality
Licenceglm-4commercial use with conditions
History
10 versionsBenchmarks
all categoriesChat & Knowledge
Arena Text · overall1,273#169
Arena Text · Chinese1,228.6#166
Arena Text · English1,241.9#176
Arena Text · German1,183.9#147
Arena Text · instruction following1,202.7#177
Arena Text · Korean1,042#148
Arena Text · non-English1,199.6#176
Arena Text · Russian1,213.1#174
Arena Text · overall · October 22, 20261,225.9#170
Writing & Language
Arena Text · creative writing1,241#164
FAQ
How much does GLM-4 cost?›
There is no per-unit API price: price not verified.
How good is GLM-4, and what is it best at?›
#176 of 205 in Chat & Knowledge, #171 of 197 in Writing & Language.
Is GLM-4 better than Qwen2.5 Coder 32B Instruct?›
On Arena Text · English, GLM-4 scores 1,241.9 vs 1,247.8 — about the same.
When was GLM-4 released? Is there a newer version?›
Released 2024-06-04. It is the current version in its line.
Is GLM-4 open source?›
Yes, the weights are open (glm-4 licence).
Can I run GLM-4 locally?›
The weights are open and can be self-hosted; hardware needs are not confirmed yet.
Can GLM-4 read images?›
No, it accepts text only.
What is GLM-4?›
GLM-4 is an open-weights language model from Z.ai, released Jun 2024.
What are the best alternatives to GLM-4?›
GPT-5 Mini (better), NVIDIA Nemotron 3.5 Lightning 30B A3B NVFP4 (open weights).