Kimi K3 models
3 Kimi K3 models from Moonshot AI on Hugging Face, from 2.2B to 3.6B parameters.
Part of the Kimi series · Previous generation: Kimi K2
Fine-tunes and community models (3)
| Model | Parameters | Published as | VRAM needed | Live GPU fit | Est. $/hr |
|---|---|---|---|---|---|
| Kimi-K3-DSpark | 2.2B | BF16 | 5.0 GB | RTX 3070 | $0.088/hr |
| Kimi-K3-DFlash | 2.6B | BF16 | 5.8 GB | RTX 3070 | $0.088/hr |
| Kimi-K3-DSpark | 3.6B | BF16 | 8.0 GB | RTX 3070 | $0.088/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.
Run and fine-tune Kimi K3
- How to run Kimi K3: engine support, chat template and context flags, with a serving command sized per model.
- How to fine-tune Kimi K3: LoRA and QLoRA memory per model, and the license limits on derivatives.
More from Moonshot AI
- Kimi K2 (5 models)
- Kimi VL (1 models)
- moonshotai (1 models)
- Moonlight (2 models)