ibm-research models
2 ibm-research models from ibm-research on Hugging Face, from 3.4B to 3.5B parameters, published by ibm-research in F32, with a 4K-token context. The smallest official model, PowerMoE-3b, needs about 15.1 GB of VRAM at its published precision; the cheapest live fit is V100 at $0.110/hr.
Pick a size
One row per official ibm-research 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 |
|---|---|---|---|---|---|---|---|
| PowerMoE-3b | 3.4B | 15.1 GB | 3.8 GB | 1.9 GB | V100 | $0.110/hr | 2.0 GB |
| PowerLM-3b | 3.5B | 15.7 GB | 3.9 GB | 2.0 GB | V100 | $0.110/hr | 11.3 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 (2)
| Model | Parameters | Published as | VRAM needed | Live GPU fit | Est. $/hr |
|---|---|---|---|---|---|
| PowerMoE-3b | 3.4B | F32 | 15.1 GB | V100 | $0.110/hr |
| PowerLM-3b | 3.5B | F32 | 15.7 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.