EXAONE 3.5 models
4 EXAONE 3.5 models on Hugging Face, from 2.4B to 32.0B parameters. At the precision each one is published in, the smallest needs about 10.8 GB of VRAM (EXAONE-3.5-2.4B-Instruct, cheapest live fit: V100) and the largest about 143 GB (EXAONE-3.5-32B-Instruct, cheapest live fit: V100). The cheapest way to run EXAONE-3.5-2.4B-Instruct is $0.088/hr.
EXAONE 3.5 models
| Model | Parameters | Published as | VRAM needed | Live GPU fit | Est. $/hr |
|---|---|---|---|---|---|
| EXAONE-3.5-2.4B-Instruct | 2.4B | F32 | 10.8 GB | V100 | $0.088/hr |
| EXAONE-3.0-7.8B-Instruct | 7.8B | F32 | 35.0 GB | RTX A6000 | $0.363/hr |
| EXAONE-3.5-7.8B-Instruct | 7.8B | F32 | 35.0 GB | RTX A6000 | $0.363/hr |
| EXAONE-3.5-32B-Instruct | 32.0B | F32 | 143 GB | V100 × 5 | $0.935/hr |
VRAM is for the precision the model is published in, with the same overhead and an 8,192-token context assumed on every page; 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.
More from LG AI Research
- EXAONE 4.0 (3 models)
