Raon 1 nano
Raon 1 nano is an open-weights language model from KRAFTON, released October 29, 2026. It has 3.5B parameters (dense). The context window is 256K tokens.
Check it against your own PC →
Lab
KRAFTON
Released
October 29, 2026
Weights
Open apache-2.0
Parameters
3.5B dense
Context
256K
Price / M tokens
–
Epoch ECI
not yet rated
Can you run Raon 1 nano locally?
The smallest common GPU it runs on is 6 GB of VRAM (with 32 GB of RAM): Q6_K, fully on GPU, roughly 38 tokens per second. Estimates assume 32 GB of DDR5 RAM, a 16K context and llama.cpp-style GGUF quantization. Enter your exact hardware in the live app.
| GPU memory | Verdict | Best quant | VRAM | RAM | tok/s |
|---|---|---|---|---|---|
| 6 GB VRAM | Runs on GPU | Q6_K (≈ lossless) | 3.9 GB | – | ~38 |
| 8 GB VRAM | Runs on GPU | Q6_K (≈ lossless) | 3.9 GB | – | ~53 |
| 12 GB VRAM | Runs on GPU | Q6_K (≈ lossless) | 3.9 GB | – | ~69 |
| 16 GB VRAM | Runs on GPU | Q6_K (≈ lossless) | 3.9 GB | – | ~94 |
| 24 GB VRAM | Runs on GPU | Q6_K (≈ lossless) | 3.9 GB | – | ~160 |
| 32 GB VRAM | Runs on GPU | Q6_K (≈ lossless) | 3.9 GB | – | ~270 |
Ready-made quantizations on Hugging Face: KRAFTON/Raon-1-nano-gguf · webmp3/Sakura-Raon-1-nano-GGUF · KRAFTON/Raon-1-nano-mlx · KRAFTON/Raon-1-nano-litert-lm
Benchmark results
Epoch AI has not published results for Raon 1 nano yet. New models are usually evaluated within days to weeks.