Nex-N2 — local AI model by Nex-AGI
Written by Jakub Rusinowski · Last updated
Nex-AGI's Apache-2.0 agentic pair, post-trained on Qwen3.5 and built around what Nex calls Agentic Thinking: Adaptive Thinking lets the model choose when and how deeply to reason, Coherent Thinking keeps one reasoning style across chat, tool calls and terminal work. Pro reaches 75.3 on Terminal-Bench 2.1 and 90.7 on GPQA Diamond. The 35B mini is the size that matters for local use — 21.9 GB at Q4_K_M with 3B active parameters per token. This is the generation whose parameter counts Nex states outright.
Variants
The smallest Nex-N2 variant needs about 22 GB of VRAM at Q4_K_M — quantized weights plus framework overhead, before any KV cache.
| Model | VRAM at Q4 | VRAM | Context | Run it |
|---|---|---|---|---|
| Nex-N2 mini → 35B (3B active) | ~21.9 GB | 262,144 | ollama run nex-n2 | |
| Nex-N2 Pro → 397B (17B active) | ~240.5 GB | 262,144 | ollama run nex-n2 |
Memory is quantized weights plus overhead at Q4_K_M, from the same engine as the GPU & VRAM checker.
How to run Nex-N2 locally
Install Ollama, then pull the tag.
ollama run nex-n2Pick a size above for its own VRAM figure, speed estimate and install command.
Licence
Commercial use permitted. No usage restrictions beyond attribution.
Applies to: Nex-N2 mini, Nex-N2 ProRecommended GPU
The cheapest catalogued GPU that runs Nex-N2 locally (min 22 GB VRAM) is the NVIDIA GeForce RTX 5090 (32 GB).
Can I run Nex-N2 on my GPU?
- Nex-N2 on AMD Radeon RX 7800 XT
- Nex-N2 on AMD Radeon RX 7900 XT
- Nex-N2 on AMD Radeon RX 7900 XTX
- Nex-N2 on AMD Radeon RX 9060 XT 16GB
- Nex-N2 on AMD Radeon RX 9070
- Nex-N2 on AMD Radeon RX 9070 XT
- Nex-N2 on Apple M1
- Nex-N2 on Apple M1 Pro
- Nex-N2 on Apple M2
- Nex-N2 on Apple M2 Pro
- Nex-N2 on Apple M3
- Nex-N2 on Apple M3 Pro
- Nex-N2 on Apple M3 Ultra
- Nex-N2 on Apple M4
- Nex-N2 on Apple M4 Pro
- Nex-N2 on Apple M5
- Nex-N2 on Intel Arc B570
- Nex-N2 on Intel Arc B580
- Nex-N2 on NVIDIA GeForce RTX 3060 (12GB)
- Nex-N2 on NVIDIA GeForce RTX 3080 (10GB)
- Nex-N2 on NVIDIA GeForce RTX 3090
- Nex-N2 on NVIDIA GeForce RTX 4060 Ti 16GB
- Nex-N2 on NVIDIA GeForce RTX 4070
- Nex-N2 on NVIDIA GeForce RTX 4070 Super
- Nex-N2 on NVIDIA GeForce RTX 4070 Ti
- Nex-N2 on NVIDIA GeForce RTX 4070 Ti Super
- Nex-N2 on NVIDIA GeForce RTX 4080
- Nex-N2 on NVIDIA GeForce RTX 4080 Super
- Nex-N2 on NVIDIA GeForce RTX 4090
- Nex-N2 on NVIDIA GeForce RTX 5060 Ti 16GB
- Nex-N2 on NVIDIA GeForce RTX 5070
- Nex-N2 on NVIDIA GeForce RTX 5070 Ti
- Nex-N2 on NVIDIA GeForce RTX 5080
- Nex-N2 on NVIDIA GeForce RTX 5090
- Nex-N2 on NVIDIA L40S
- Nex-N2 on NVIDIA RTX 6000 Ada Generation
Nex-N2 — frequently asked questions
How much VRAM does Nex-N2 need?
Nex-N2 needs about 22 GB VRAM at Q4_K_M quantization for its smallest variant. Variants: Nex-N2 mini (22 GB, Q4_K_M); Nex-N2 Pro (240 GB, Q4_K_M). On Apple Silicon, unified memory counts toward this requirement.
Can I run Nex-N2 on an RTX 4090 (24 GB)?
Yes — Nex-N2 runs on an RTX 4090 (24 GB) and other 24 GB cards such as the RTX 3090. Smaller variants also fit comfortably on 8–16 GB GPUs at Q4_K_M.
What quantization should I use for Nex-N2?
Q4_K_M is the best balance of quality and VRAM for Nex-N2 in most cases. Choose Q8_0 for near-lossless quality if you have spare VRAM, or smaller quants (Q3/Q2) only when memory is tight.
How do I run Nex-N2 with Ollama?
Nex-N2 has no local Ollama tag — the published tag is cloud-hosted, so running it sends your prompts to a hosted GPU rather than your own machine.