GPT-Neo models
3 GPT-Neo models from EleutherAI on Hugging Face, from 150M to 2.7B parameters, published by EleutherAI in F32, with a 2K-token context. The smallest official model, gpt-neo-125m, needs about 0.7 GB of VRAM at its published precision; the cheapest live fit is RTX 3070 at $0.088/hr.
Pick a size
One row per official GPT-Neo size: the VRAM it needs at each precision, the cheapest GPU that holds it today, and the KV cache for a 32K-token context.
| Model | Parameters | Native VRAM | FP8 VRAM | INT4 VRAM | Live GPU fit (native) | Est. $/hr | KV cache at 32K |
|---|---|---|---|---|---|---|---|
| gpt-neo-125m | 150M | 0.7 GB | 0.2 GB | 0.1 GB | RTX 3070 | $0.088/hr | 1.1 GB |
| gpt-neo-1.3B | 1.4B | 6.1 GB | 1.5 GB | 0.8 GB | RTX 3070 | $0.088/hr | 6.0 GB |
| gpt-neo-2.7B | 2.7B | 12.2 GB | 3.0 GB | 1.5 GB | V100 | $0.110/hr | 10.0 GB |
VRAM is weights times a flat 1.2 overhead; the KV cache is a separate per-model figure at 16-bit, shown where the architecture is published. See the methodology.
Official models (3)
| Model | Parameters | Published as | VRAM needed | Live GPU fit | Est. $/hr |
|---|---|---|---|---|---|
| gpt-neo-125m | 150M | F32 | 0.7 GB | RTX 3070 | $0.088/hr |
| gpt-neo-1.3B | 1.4B | F32 | 6.1 GB | RTX 3070 | $0.088/hr |
| gpt-neo-2.7B | 2.7B | F32 | 12.2 GB | V100 | $0.110/hr |
VRAM is for the precision the model is published in: the weights times a flat 1.2 overhead, with the KV cache not included. See the methodology. The fit shown is the lowest-priced single GPU type that holds the model at that precision, or the lowest-priced multi-GPU set (up to 8) when none does.