Agents

Check model cardOpen weightsUpdated June 2026Legacy baseline

GLM-4.5

Earlier GLM open-weight MoE model useful as a baseline for agent, coding, and reasoning workflows.

Z.ai · GLM

Editorial review

Reviewed byOpenSourcesAI EditorialLast updatedJune 2026SourcesZ.ai GLM-4.5 docs, Hugging Face

Model checkpoints, context windows, provider support, local runtime compatibility, and license terms can change quickly. Verify the exact model card before production or commercial use.

Best for

Teams comparing prior GLM releases with current GLM-5.1 behavior.

Who should use it

  • Teams comparing prior GLM releases with current GLM-5.1 behavior.
  • Teams with access to hosted inference or server-class deployment paths.
  • Developers evaluating coding assistant, repo-editing, and code review workflows.
  • Teams testing tool-use, agentic planning, and multi-step workflow behavior.

Common workflows

  • Agentic apps, coding, tool use, reasoning
  • agents workflows
  • coding workflows
  • reasoning workflows
  • legacy workflows

Deployment and hardware notes

Full model is server-class; Air or quantized builds are more practical.

License and usage notes

Check model card. Open weights. Verify the exact model card and license terms for the checkpoint or hosted provider you use.

Strengths

  • Open weights model option for GLM workflows.
  • Teams comparing prior GLM releases with current GLM-5.1 behavior.
  • Tracked as Legacy baseline in the OpenSourcesAI model directory.

Limitations

  • Demote behind GLM-5.1 for current 2026 coverage; production fit depends on serving support and license review.
  • Full model is server-class; Air or quantized builds are more practical.
  • Context window and limits: 128k noted in Z.ai docs.
  • Verify the exact model card, provider docs, license, and serving support before production use.

Will GLM-4.5 run on your machine?

GLM-4.5 is 355B parameters and needs 217.5 GB of VRAM at Q4_K_M216 GB of weights plus 1.5 GB of runtime overhead for the inference server itself.

VRAM by quantization

QuantizationWeightsNeeds (with overhead)Quality
Q4_K_M216 GB217.5 GBgood
Q8_0381 GB382.5 GBhigh

Fit on common hardware at Q4_K_M

HardwareMemory the model can useSystem RAMVerdict
CPU OnlyNone (CPU only)16 GBToo large
RTX 4060 Laptop8 GB16 GBToo large
RTX 3060 (12GB)12 GB32 GBToo large
RTX 4060 Ti (16GB)16 GB32 GBToo large
RTX 309024 GB64 GBToo large
Apple Silicon (Unified Memory) 36 GB27 GB of 36 GB36 GBToo large
RTX 509032 GB64 GBToo large
Apple Silicon (Unified Memory) 48 GB36 GB of 48 GB48 GBToo large
Apple Silicon (Unified Memory) 64 GB48 GB of 64 GB64 GBToo large
Apple Silicon (Unified Memory) 96 GB72 GB of 96 GB96 GBToo large
Apple Silicon (Unified Memory) 128 GB96 GB of 128 GB128 GBToo large
Apple Silicon (Unified Memory) 192 GB144 GB of 192 GB192 GBToo large

Comfortable means VRAM clears the requirement by 2 GB or more. Tight means it covers the requirement with no margin. CPU offload means the model does not fit in VRAM but system RAM is at least 1.6× the weights, so it will run at reduced speed — expect roughly 1–5 tokens per second. Figures are weights plus a fixed runtime overhead and exclude KV-cache growth, which scales with context length.

Apple Silicon shares one pool of memory between the system and the GPU, so a model cannot use all of it. These rows apply the same 75% usable fraction the Compatibility Checker uses, which is why a 36 GB Mac is graded on less than 36 GB.

VRAM fit by quantization level

Enter your GPU VRAM below to see which quantization of GLM-4.5 fits and get the Ollama run command.

Related resources

Continue with model source notes, local tools, and implementation guides related to this model.

HardwareServer-classRuntimevLLM, SGLang, Z.ai API, hosted providersContext128k noted in Z.ai docsLast updated2026
Z.ai GLM-4.5 docs

Model ecosystem connections

Use these next-step links to move from this profile into related tools, comparisons, guides, stacks, and curated shortlists.