Cerebras GPT 13B — Hardware Requirements & GPU Compatibility
Chat
Cerebras GPT 13B is a 13B-parameter open language model from Cerebras. It supports a context window of up to 2,048 tokens. At BF16 it needs about 28.60 GB of VRAM — see which GPUs and Macs can run it below.
Doesn't fit a 24 GB GPU
Smallest consumer fit → RTX 5090 · BF16 · 28.6 GB · ~41 tok/s
Macs from 36 GB (BF16 · ~11 tok/s)
Apache 2.0
Not set up for it?
Reference GPU · RTX 4090 24 GB
Estimates, ±20% · how we estimate
Run Cerebras GPT 13B
One recommended way. Every other method, OS and quant is in the step-by-step guide.
Recommended
vllm serve cerebras/Cerebras-GPT-13B --max-model-len 2048vLLM repo with its own quantisation
Download: Weights on Hugging Face
Other ways
each opens step-by-stepRun in cloud
Fits on RTX A6000 (48 GB) (19 GB headroom) · BF16
- Generation speed
- ~18 tok/s
- generation speed
- Cost per 1M output tokens
- $5.24
- per 1M output tokens
How Much VRAM Does Cerebras GPT 13B Need?
Select a quantization level; the listed hardware adheres to it.
| Quantization | VRAM | +32K | On RTX 4090 (reference) | |
|---|---|---|---|---|
| 28.6 GB | — | 16-bit · bfloat — Brain floating point 16 — preferred for training | Does not fit: needs 28.6 GB |
Selected BF16
Brain floating point 1628.6GBat 2K context
- Weights
- 27.9 GB
- KV cache
- 0.4 GB
- Overhead
- 0.3 GB
Doesn't fit RTX 4090 (reference) · needs 4.6 GB more
≥30 tok/s needs ≥1.32 TB/s · RTX 5090 ~41
Which Hardware Can Run Cerebras GPT 13B?
BF16 · 28.6 GBCerebras GPT 13B (BF16) requires 28.6 GB of VRAM to load the model weights. For comfortable inference with headroom for KV cache and system overhead, 38+ GB is recommended. 1 GPU can run it, including NVIDIA GeForce RTX 5090. 78 devices with unified memory can run Cerebras GPT 13B, including Mac Studio (M3 Ultra, 256GB), Mac Studio (M3 Ultra, 512GB), Mac Studio M5 Max (36 GB).
32–40 GB
fits · 3.4 GB free · up to 2K ctx
1 GPU · 2 Mac classes · 1 device family
RTX 5090~41Simulate output on RTX 5090M5 Max36 GB~11Simulate output on Mac Studio M5 Max (36 GB)M3 Pro*36 GB~4Simulate output on MacBook Pro 14" M3 Pro (36 GB)Orin~5Simulate output on NVIDIA Jetson AGX Orin 32GB
By VRAM tier · typical GPUs and Macs at BF16
How good is Cerebras GPT 13B?
llmrun Score blends public benchmark boards into one number; 100 = the average open model.
Not enough data yet for an llmrun Score.
A Score needs results on enough independent boards. How the Composite Score is built
Benchmarks
The 6 boards that weigh most in the llmrun Score, of 7 with a result for this model; badges show who ran the test.
- MMLU
Epoch AI · Mar 2023Lab-reported
Lab-reported26.2%
#82/85 open
- HellaSwag
Epoch AI · Mar 2023Lab-reported
Lab-reported59.4%
#40/47 open
- ARC Challenge
Epoch AI · Mar 2023Lab-reported
Lab-reported32.4%
#53/56 open
- WinoGrande
Epoch AI · Mar 2023Lab-reported
Lab-reported60.8%
#41/51 open
- BoolQ
Epoch AI · Mar 2023Lab-reported
Lab-reported61.1%
#37/38 open
- OpenBookQA
Epoch AI · Mar 2023Lab-reported
Lab-reported35.8%
#23/26 open
Similar models
6| Model | Score | VRAM | tok/s | Why it is similar |
|---|---|---|---|---|
| Gemma 3 12B ITGoogle | 74.2 | 10.9 GB | ~82 | same size |
| Mistral Nemo Base 2407Mistral AI | — | 11 GB | ~81 | same size |
| Phi 3 Medium 128k InstructMicrosoft | — | 12 GB | ~72 | same size |
| Baichuan2 13B Basebaichuan-inc | — | 11.5 GB | ~76 | same size |
| Llama 2 13B Chat HFMeta | — | 11.5 GB | ~76 | same size |
| Vicuna 13B V1.3LMSYS | — | 8.6 GB | ~76 | same size |
Specifications
- Publisher
- Cerebras
- Parameters
- 13B
- Architecture
- GPT2Model
- Context Length
- 2,048 tokens
- Vocabulary Size
- 50,257
- Release Date
- 2023-03-20
- License
- Apache 2.0
Frequently Asked Questions
How much VRAM does Cerebras GPT 13B need?
Cerebras GPT 13B requires 28.6 GB of VRAM at BF16.
VRAM = Weights + KV Cache + Overhead
Weights = 13B × 16 bits ÷ 8 = 26 GB
KV Cache + Overhead ≈ 2.6 GB (at 2K context + ~0.3 GB framework)
Fit ratings and hardware model lists check this model with room for a 16K-token context, which needs a little more memory.
VRAM usage by quantization
Can I run Cerebras GPT 13B on a Mac?
Yes — Mac Studio M5 Max (36 GB) and 65 other Macs can run Cerebras GPT 13B. Apple Silicon uses unified memory, so the model shares RAM with the system. At BF16 you need at least 28.6 GB of usable unified memory (RAM minus macOS overhead).
Can I run Cerebras GPT 13B locally?
Yes — Cerebras GPT 13B can run locally on consumer hardware. Popular tools include Ollama, LM Studio, and llama.cpp.
How fast is Cerebras GPT 13B?
At BF16, Cerebras GPT 13B can reach ~168 tok/s on AMD Instinct MI350X. Speed depends mainly on GPU memory bandwidth. Real-world results typically within ±20%.
tok/s = (bandwidth GB/s ÷ model GB) × efficiency
Example: NVIDIA B200 → 8000 ÷ 28.6 × 0.65 = ~182 tok/s
Estimated speed at BF16 (28.6 GB)
Real-world results typically within ±20%. Speed depends on batch size, quantization kernel, and software stack.
What's the download size of Cerebras GPT 13B?
At BF16, the download is about 26.00 GB.
Which GPUs can run Cerebras GPT 13B?
1 consumer GPU can run Cerebras GPT 13B at BF16. Top options include NVIDIA GeForce RTX 5090.
Which devices can run Cerebras GPT 13B?
78 devices with unified memory can run Cerebras GPT 13B at BF16 (28.6 GB), including ASUS Ascent GX10, Beelink GTR9 Pro (Ryzen AI Max+ 395, 128 GB), Framework Desktop (Ryzen AI Max+ 395, 128 GB), GMKtec EVO-X2 (Ryzen AI Max+ 395, 128 GB). Apple Silicon Macs use unified memory shared between CPU and GPU, making them well-suited for local LLM inference.