Best GPU for GLM-4.5: VRAM Requirements
GLM-4.5 needs about 784.4 GB of GPU memory in BF16 and 198.6 GB at 4 bits. See which GPUs fit it for inference and fine-tuning.
GLM-4.5 needs about 784.4 GB of VRAM in FP16, 393.9 GB in FP8 and 198.6 GB at 4 bits for inference with an 8,192-token context. In FP16 it fits on 4 × NVIDIA B300 288GB; a 4-bit build fits on one NVIDIA B300 288GB.
GLM-4.5 is a mixture-of-experts language model with 355B total and 32B active parameters per token from Z.ai, released in July 2025, with a 131,072-token context window and published under the MIT. Its weights alone take 710 GB in BF16 and 177.5 GB at 4 bits per parameter; the tables below add the KV cache and runtime memory, then list the GPUs that hold each setup.
Recommended GPUs for GLM-4.5
Best for inference
2 GPUs × NVIDIA B300 288GB
FP8 or INT8 weights, 393.9 GB at 8,192 tokens.
Best for fine-tuning
4 GPUs × NVIDIA B300 288GB
LoRA on BF16 weights, 923 GB.
Budget pick
4 GPUs × NVIDIA RTX PRO 6000 Blackwell 96GB
Smallest card class that holds a 4-bit build, 198.6 GB.
GLM-4.5 VRAM requirements by precision
One request with an 8,192-token context, KV cache in BF16, plus a 10 % margin for activations and runtime buffers.
| Precision | Weights | KV cache | Total for inference |
|---|---|---|---|
| BF16 / FP16 | 710 GB | 3.1 GB | 784.4 GB |
| FP8 | 355 GB | 3.1 GB | 393.9 GB |
| INT4 | 177.5 GB | 3.1 GB | 198.6 GB |
GLM-4.5 VRAM by task
Fine-tuning rows assume a short sequence, batch 1 and gradient checkpointing; longer sequences add activation memory.
| Task | GPU memory |
|---|---|
| Inference at the full 131,072-token context (FP8) | 444.8 GB |
| QLoRA (4-bit base) | 230.8 GB |
| LoRA (BF16 base) | 923 GB |
| Full fine-tuning (BF16, AdamW) | 6,248 GB |
GPUs that fit GLM-4.5
Number of cards whose combined memory holds each setup, assuming the model is split across them with tensor or pipeline parallelism. A dash means more than eight cards.
| GPU | Memory | BF16 inference (784.4 GB) | FP8 inference (393.9 GB) | INT4 inference (198.6 GB) | QLoRA (230.8 GB) | LoRA (923 GB) |
|---|---|---|---|---|---|---|
| NVIDIA GeForce RTX 5090 32GB | 32 GB | — | — | 8 GPUs | 8 GPUs | — |
| NVIDIA Tesla V100 32GB | 32 GB | — | — | 8 GPUs | 8 GPUs | — |
| NVIDIA A100 PCIe 40GB | 40 GB | — | — | 8 GPUs | 8 GPUs | — |
| NVIDIA L40S 48GB | 48 GB | — | — | 8 GPUs | 8 GPUs | — |
| NVIDIA Quadro RTX 8000 48GB | 48 GB | — | — | 8 GPUs | 8 GPUs | — |
| NVIDIA RTX 6000 Ada 48GB | 48 GB | — | — | 8 GPUs | 8 GPUs | — |
| NVIDIA RTX A6000 48GB | 48 GB | — | — | 8 GPUs | 8 GPUs | — |
| NVIDIA A100 PCIe 80GB | 80 GB | — | 8 GPUs | 4 GPUs | 4 GPUs | — |
| NVIDIA A100 SXM 80GB | 80 GB | — | 8 GPUs | 4 GPUs | 4 GPUs | — |
| NVIDIA H100 PCIe 80GB | 80 GB | — | 8 GPUs | 4 GPUs | 4 GPUs | — |
| NVIDIA H100 SXM 80GB | 80 GB | — | 8 GPUs | 4 GPUs | 4 GPUs | — |
| NVIDIA RTX PRO 6000 Blackwell 96GB | 96 GB | — | 8 GPUs | 4 GPUs | 4 GPUs | — |
| NVIDIA H200 141GB | 141 GB | 8 GPUs | 4 GPUs | 2 GPUs | 2 GPUs | 8 GPUs |
| NVIDIA B200 180GB | 180 GB | 8 GPUs | 4 GPUs | 2 GPUs | 2 GPUs | 8 GPUs |
| NVIDIA B300 288GB | 288 GB | 4 GPUs | 2 GPUs | 1 GPU | 1 GPU | 4 GPUs |
| NVIDIA GB300 NVL | 288 GB | 4 GPUs | 2 GPUs | 1 GPU | 1 GPU | 4 GPUs |
What is GLM-4.5?
GLM-4.5 is a mixture-of-experts model from Z.ai built for intelligent agents, with 355 billion total and 32 billion active parameters. It is a hybrid reasoning model, with a thinking mode for complex reasoning and tool use and a non-thinking mode for immediate answers.
- Developer
- Z.ai
- Released
- July 2025
- Parameters
- 355B total, 32B active per token
- Context window
- 131,072 tokens
- License
- MIT
The published checkpoint also holds a multi-token prediction layer (about 358 billion parameters in all), not counted here. Z.ai also publishes FP8 weights.
Source: GLM-4.5 model card (as of November 17, 2026).
Frequently asked questions
- How much VRAM does GLM-4.5 need?
- For inference with an 8,192-token context and one request, GLM-4.5 needs about 784.4 GB in BF16, 393.9 GB in FP8, 198.6 GB in INT4, including a 10 % runtime margin. Real usage depends on the inference engine and its settings.
- Can GLM-4.5 run on a single GPU?
- Yes. In INT4 it needs about 198.6 GB, which fits one NVIDIA B300 288GB (288 GB). Higher precisions need a larger card or several GPUs.
- How much GPU memory does fine-tuning GLM-4.5 take?
- QLoRA (4-bit base): about 230.8 GB; LoRA (BF16 base): about 923 GB; Full fine-tuning (BF16, AdamW): about 6,248 GB. Longer training sequences and larger batches add activation memory on top.
- What is the best GPU for GLM-4.5?
- Best for inference: 2 GPUs × NVIDIA B300 288GB (FP8 or INT8 weights, 393.9 GB at 8,192 tokens). Best for fine-tuning: 4 GPUs × NVIDIA B300 288GB (LoRA on BF16 weights, 923 GB). Budget pick: 4 GPUs × NVIDIA RTX PRO 6000 Blackwell 96GB (Smallest card class that holds a 4-bit build, 198.6 GB). Check live offers before booking, as availability changes.
- Can GLM-4.5 run on an RTX 4090?
- No: even a 4-bit build needs about 198.6 GB, more than eight RTX 4090 cards hold (192 GB). It needs data center GPUs with more memory per card.
- Can GLM-4.5 run on an RTX 5090?
- Not on one card: even a 4-bit build needs about 198.6 GB, so it takes 8 RTX 5090 cards (256 GB combined) with the model split across them.
Estimates use the formulas of the LLM VRAM calculator (decimal GB). Real usage depends on the inference engine, quantization format and memory fragmentation: test on the target GPU before a long rental.
Explore the live market
Every offer on the live market comes from a reviewed supplier and is sold by EverXCloud Inc.