Axion 1.0 35B
Axion 1.0 35B is an open-weights language model from Axion1-0, released October 16, 2026. It has 35B parameters (mixture-of-experts, 3.9B active). The context window is 256K tokens.
Check it against your own PC →
Can you run Axion 1.0 35B locally?
The smallest common GPU it runs on is 6 GB of VRAM (with 32 GB of RAM): Q6_K, GPU + experts in RAM, roughly 19 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 · GPU + CPU experts | Q6_K (≈ lossless) | 5.2 GB | 24.5 GB | ~19 |
| 8 GB VRAM | Runs · GPU + CPU experts | Q6_K (≈ lossless) | 7.2 GB | 22.5 GB | ~22 |
| 12 GB VRAM | Runs · GPU + CPU experts | Q6_K (≈ lossless) | 11.2 GB | 18.5 GB | ~27 |
| 16 GB VRAM | Runs · GPU + CPU experts | Q6_K (≈ lossless) | 15.2 GB | 14.5 GB | ~34 |
| 24 GB VRAM | Runs on GPU | Q4_K_M (very good) | 22.4 GB | – | ~190 |
| 32 GB VRAM | Runs on GPU | Q6_K (≈ lossless) | 29.7 GB | – | ~250 |
Ready-made quantizations on Hugging Face: prithivMLmods/Axion-1.0-35B-GGUF · mradermacher/Axion-1.0-35B-GGUF · mradermacher/Axion-1.0-35B-i1-GGUF · IsValorum/Axion-1.0-35B-APEX-I-NanoPlus-V2.1-GGUF · IsValorum/Axion-1.0-35B-APEX-I-MiniPlus-V2.1-GGUF · IsValorum/Axion-1.0-35B-Q8_0-GGUF
Benchmark results
Epoch AI has not published results for Axion 1.0 35B yet. New models are usually evaluated within days to weeks.