CohereLabs models
3 CohereLabs models from Cohere on Hugging Face, from 2.1B to 30.5B parameters, published by Cohere in BF16, with a 500,000-token context. The smallest official model, cohere-transcribe-03-2026, needs about 4.6 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 CohereLabs 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 |
|---|---|---|---|---|---|---|---|
| cohere-transcribe-03-2026 | 2.1B | 4.6 GB | 2.3 GB | 1.2 GB | RTX 3070 | $0.088/hr | not published for this architecture |
| North-Mini-Code-1.0 | 30.5B | 68.1 GB | 34.1 GB | 17.0 GB | A100 | $1.31/hr | 1.1 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 |
|---|---|---|---|---|---|
| cohere-transcribe-03-2026 | 2.1B | BF16 | 4.6 GB | RTX 3070 | $0.088/hr |
| cohere-transcribe-arabic-07-2026 | 2.1B | BF16 | 4.6 GB | RTX 3070 | $0.088/hr |
| North-Mini-Code-1.0 | 30.5B | BF16 | 68.1 GB | A100 | $1.31/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.