Ctrl K
Back to models

Z.AI

GLM-5.3

  • Released
  • Open weights
  • GLM-5.3 License
Compare
Intelligence Index

Max setting

44.8#13 of 83measured models
Measured · Artificial Analysis v4.3.2
API price from
$1.40 in / $4.40 out / 1M tokZ.AI · direct API
Context window
1.05M tok

Performance profile

ReasoningRank #14 of 107
91.7%
GPQA Diamond
CodingRank #10 of 94
59.0%
SciCode
Cost efficiencyRank #56 of 67
22.3pts / $
Intelligence ÷ eval cost

Capabilities

Context window

Modalities

Reasoning modes

Avg reasoning tokens / task — Artificial Analysis Intelligence Index v4.3.2: weighted average reasoning tokens per benchmark task, not a token limit. Source (opens in a new tab)

Reasoning trade-offs

Source (opens in a new tab)
Intelligence Index v4.3.2
GLM-5.3
Cost per task (USD) · log scale
GLM-5.3Max
Intelligence Index
44.8
Cost / task
$2.01
Evaluation time
12.9m

Pricing & access

API route
Z.AIAPI
Tier & contextStandard

USD per 1M tokens

Input
$1.40
Output
$4.40
Cached input
$0.26

Run locally

GGUFQ8_0
Context
Estimated memory

~821.8 GiB

RAM + VRAM

Estimate assumptions

One text chat at the selected context, with a 16-bit cache and app allowance. Operating system and image processing are extra; optimized caches can use less.

1024 GiB RAMwithout GPU offload

Memory budgetOne text chat
Weights
746.3 GiB
Context cache
0.8 GiB16-bit cache
App allowance
74.6 GiB
Download
801.4 GB17 files · keep together
Disk space
~887 GB
Available software
Ollama (opens in a new tab)LM Studio (opens in a new tab)Jan (opens in a new tab)KoboldCpp (opens in a new tab)llama.cpp (opens in a new tab)
Get GGUF (opens in a new tab)
Set up in Ollama · Q8_0 · 8K context

Save the following as Modelfile. The context matches the estimate above.

FROM hf.co/unsloth/GLM-5.3-GGUF:Q8_0
PARAMETER num_ctx 8192

Then run these commands in your terminal:

ollama create llmcurrent-local -f Modelfile
ollama run llmcurrent-local

Ollama import instructions (opens in a new tab) · For LM Studio or Jan, open the app’s model search, enter unsloth/GLM-5.3-GGUF, choose Q8_0, and set context to 8,192 tokens in the app’s model settings. Software links above open each app’s setup guide.

vLLM setup (opens in a new tab)

Benchmarks

Full leaderboardSource (opens in a new tab)

Humanity’s Last Exam

Frontier reasoning

GLM-5.3 · Max#23 of 126

  1. 1Claude Opus 5.5Max61.4%
  2. 2Claude Fable 5.1Max59.1%
  3. 3Gemini 4 ArgonHigh57.1%
  5. 20GPT-5.6 TerraMax42.9%
  6. 21GPT-5.3-CodexExtra high42.5%
  7. 22Qwen3.8 2.4T A95BReasoning42.4%
  8. 23GLM-5.3Max42.3%
  9. 24DeepSeek-V4-Pro-0813Max41.0%
  10. 25GLM-5.3-FlashMax39.9%
  11. 26Ling-3.1-flashReasoning39.4%

Peers use their best measured result on this benchmark.

Alternatives

Each model at its highest-scoring measured setting · Intelligence Index v4.3.2 and its evaluation cost per task.