What GPU do I need to run zai-org/GLM-5.3-BF16?
753.3B parameters, published in BF16. View on Hugging Face
Set up GLM-5.3-BF16GLM-5.3-BF16 is published by zai-org on Hugging Face, with 4,111 downloads and 30 likes to date. It's a GlmMoeDsaForCausalLM model built for text-generation, published natively in BF16.
VRAM required & cheapest live GPU fit
Required VRAM = weight size at each precision times a flat 1.2 overhead for activations and fragmentation. The KV-cache grows with context and is not in that factor; it is listed per model below where the architecture is published. Full formula and assumptions: methodology.
| Precision | Weight size | Required VRAM | Cheapest live fit | GPUs needed | Est. $/hr (full fit) |
|---|---|---|---|---|---|
| BF16 | 1403.2 GB | 1683.8 GB | No capable live offer found | – | – |
| FP8 (quantized) | 701.6 GB | 841.9 GB | AMD MI300X | 5 | $13.14/hr |
| INT4 (quantized) | 350.8 GB | 421.0 GB | RTX PRO 6000 | 5 | $6.60/hr |
A GPU is only matched to a row if its hardware supports that precision, and the primary recommendation is always a single-GPU fit when one exists.
INT4 requires a quantized checkpoint actually published for this model, check its Hugging Face page before relying on this row.
GLM-5.3-BF16: KV cache by context length
The KV cache is the memory the attention layers hold for every token of context, on top of the weights. It is not part of the flat 1.2x overhead in the table above, grows with context length and with every concurrent request, and is why a long-context deployment needs more VRAM than the table shows.
Multi-head latent attention: every layer caches one compressed latent per token. Cached per token: 89,856 bytes at 16-bit.
| Context | KV cache, one sequence |
|---|---|
| 4K | 0.34 GB |
| 32K | 2.7 GB |
| 128K | 11.0 GB |
| 1M (model maximum) | 87.8 GB |
Computed from the layer, head and window counts in the model's own configuration (zai-org/GLM-5.3-BF16), at 16 bits per cached value. Each concurrent sequence needs its own cache, on top of the weights.
GLM-5.3-BF16: common questions
Can GLM-5.3-BF16 run on a single GPU?
Not on a desktop card. At BF16 it needs 1683.8 GB of VRAM, more than a single 32 GB desktop card holds. No card currently listed on the marketplace both supports BF16 and has enough VRAM for it, so how many it would take is not something this page can answer today.
Weight-to-VRAM math, the fit rules, and how live prices are normalized: full methodology.
More GLM-5 models
- GLM-5.2-FP8 (753.3B, F8_E4M3)
- GLM-5.2 (753.3B, BF16)
- GLM-5.3 (753.3B, F8_E4M3)
- penclaw-GLM-5.3-abliterated (753.3B, BF16)
- GLM-5.3-F.U-AnthraClaud-Edition-BF16 (753.3B, BF16)
Lineage
Quantized and fine-tuned versions
Alternatives at this size
Other models for text-generation within about a third of GLM-5.3-BF16's 753.3B parameters, from other model lines.
- DeepSeek-R1 (684.5B, F8_E4M3)
- NVIDIA-Nemotron-3-Ultra-550B-A55B-BF16 (560.5B, BF16)
- Kimi-K2-Instruct (1026.4B, F8_E4M3)
- A.X-K2 (691.7B, F8_E4M3)
- MiMo-V2.5-Pro (1023.2B, F8_E4M3)
More on GLM-5.3-BF16
Best chat and assistants models: how GLM-5.3-BF16 ranks against the rest.
Related reading: H100 pricing and specs, How much VRAM you need for LLMs, Serving LLMs with vLLM, vLLM vs TensorRT-LLM vs SGLang, and Best GPU for LLM inference.
