Aleph-Alpha models
2 Aleph-Alpha models from Aleph-Alpha on Hugging Face, from 78.1B to 78.1B parameters, published by Aleph-Alpha in F8_E4M3 and BF16, with a 256K-token context. The smallest official model, Kolibri-1, needs about 87.3 GB of VRAM at its published precision; the cheapest live fit is RTX PRO 6000 at $1.38/hr.
Pick a size
One row per official Aleph-Alpha 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 |
|---|---|---|---|---|---|---|---|
| Kolibri-1 | 78.1B | 87.3 GB | – | 43.6 GB | RTX PRO 6000 | $1.38/hr | 0.66 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 |
|---|---|---|---|---|---|
| Kolibri-1 | 78.1B | F8_E4M3 | 87.3 GB | RTX PRO 6000 | $1.38/hr |
| Kolibri-1-BF16 | 78.1B | BF16 | 175 GB | AMD MI300X | $2.63/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.