A.X K2
A.X K2 is an open-weights language model from SK Telecom, released July 29, 2026. It has 692B parameters (mixture-of-experts, 36B active). The context window is 256K tokens.
We introduce A.X K2, a 688B-parameter Mixture-of-Experts (MoE) language model trained from scratch as a high-performance foundation for agentic applications. A.X K2 is trained on approximately 8.5T tokens in total—fewer than its predecessor, A.X K1—by prioritizing a smaller but higher-quality data mixture; it …
Check it against your own PC →
Can you run A.X K2 locally?
Even with 32 GB of VRAM and 32 GB of RAM it is too large to run at a usable speed on a single consumer GPU. 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 | Too large | – | – | – | – |
| 8 GB VRAM | Too large | – | – | – | – |
| 12 GB VRAM | Too large | – | – | – | – |
| 16 GB VRAM | Too large | – | – | – | – |
| 24 GB VRAM | Too large | – | – | – | – |
| 32 GB VRAM | Too large | – | – | – | – |
Ready-made quantizations on Hugging Face: Baekpica/A.X-K2-VQ-GGUF · skt/A.X-K2-GGUF · skt/A.X-K2-NVFP4
Benchmark results
Epoch AI has not published results for A.X K2 yet. New models are usually evaluated within days to weeks.
Links
Announcement · Hugging Face: skt/A.X-K2