The LLMCheck index estimates GLM-5.3 at ~17 tok/s on an M5 Ultra at Q4_K_M, a figure from its published bandwidth model, not a measured run. It needs about 256 GB of unified memory. It is Zhipu AI's 753B mixture-of-experts model (40B active), released under the GLM-5.3 licence.
Want to run GLM-5.3 faster — or step up to bigger models? Find your Mac →This is an LLMCheck index estimate from its published two-term bandwidth model, not a measured run; the index labels it so everywhere it appears. Own this Mac? Submit a real run →
| Metric | Value |
|---|---|
| Tokens per second | estimated~17 tok/s |
| Provenance | estimate · how estimates are made |
| Quantization | Q4_K_M |
| Engine | MLX |
| Figure dated | 2026-09 |
| Minimum RAM | 256 GB |
| Parameters | 753B MoE (40B active) |
| Chip bandwidth | 1228 GB/s |
Each row as the index holds it, with its source. “Mac RAM” is the memory of the Mac a run used, not the model’s requirement (256 GB). The headline figure follows the reference rule: the lower-median measured plain-decode run at 5 bits or fewer (a run whose quant was not stated counts only when no stated-quant run qualifies), else the estimate; a run marked “not used to anchor estimates” is listed here but never becomes the headline.
| Speed | Quant | Engine | Mac RAM | Context | Source | Note |
|---|---|---|---|---|---|---|
| estimated27 tok/s | Q2_K | MLX | 256 GB | — | LLMCheck estimator | — |
| estimated17 tok/s | Q4_K_M | MLX | 512 GB | — | LLMCheck estimator | — |
Ollama — the index has not verified an Ollama tag for GLM-5.3, so it doesn’t print one (a guessed tag fails with “pull model manifest: file does not exist”). Search the Ollama library →
MLX — pick a build from Hugging Face’s MLX listings, then run it with its repo id:
Downloads: Search the Ollama library → · model card →
Prefer a GUI? LM Studio downloads GGUF and MLX builds in one click. Step by step: install Ollama on a Mac.
| Chip | Speed | Quant | Engine | Bandwidth |
|---|---|---|---|---|
| M3 Ultra | estimated11 tok/s | Q4_K_M | MLX | 819 GB/s |
Zhipu's post-training upgrade of GLM 5.2 on the same 753B MoE base (~40B active): Terminal-Bench 2.1 88.2, DeepSWE v1.1 66.9 (model card). Released 14 August; weights public 25 August 2026 under the custom GLM-5.3 License — MIT-like for everyone except model-as-a-service operators above US$10B revenue, who need Z.AI's security review. 2-bit is 239 GB, so it starts at a 256 GB Mac Studio (roughly 27 tok/s there); at 4-bit it needs the 512 GB model, ~17 tok/s estimated on M5 Ultra. Where GLM 5.2 stays MIT, 5.3 does not.
GLM-5.3 licence · 753B MoE · 1000K context · LLMCheck capability 50/50 · see its leaderboard rank →
GLM-5.3 needs about 256 GB of unified memory. These current Apple Silicon Macs have the headroom to run it comfortably:
As an Amazon Associate, LLMCheck earns from qualifying purchases. These affiliate links cost you nothing extra and help keep the LLMCheck index free.
See how GLM-5.3 stacks up against other models on your specific Mac hardware.
Open Compare Tool Full Leaderboard