Model library / nex-agi
Nex-N2.5-Pro is a 396.8 billion-parameter MoE language model designed for long-horizon, agentic tasks with multimodal and computer-use capabilities.
Mixture of experts, count not published
Context length in tokens, as published
Weights at the released precision
Licence declared on the repository
Overview
Nex-N2.5-Pro is part of the Nex-N2.5 family released by Nex-AGI. The publisher describes it as a next-generation agentic model built for extended tasks in real-world environments, with improvements in computer operation, web browsing and visual feedback. It aims to act continuously, self-correct, and use vision as an interface for perception and verification.
In practice the model offers 396.8 billion parameters in a Qwen3_5MoeForConditionalGeneration architecture with a mixture-of-experts design (expert count not disclosed). It supports a context window of 262,144 tokens, is distributed under the Apache-2.0 licence, and uses compressed-tensors for weight storage. These characteristics affect hardware requirements, inference latency, and licensing compliance for deployment.
Specification
| Repository | nex-agi/Nex-N2.5-Pro |
| Publisher | nex-agi |
| Published | 8 September 2026 |
| Architecture | Qwen3_5MoeForConditionalGeneration |
| Model type | qwen3_5_moe |
| Parameters | 396,802,360,816 |
| Layers | 60 |
| Hidden size | 4,096 |
| Attention heads | 32, 2 key/value heads |
| Context length | 262,144 tokens |
| Vocabulary | 248,320 tokens |
| Quantisation | compressed-tensors |
| Licence | apache-2.0 |
Memory
Parameter count multiplied by the bytes each format uses. The key-value cache comes on top.
| Precision | Weights | Notes |
|---|---|---|
| bf16 | 739.1 GiB | Full precision as released |
| fp8 | 369.6 GiB | 8-bit, near-lossless on most models |
| int4 | 184.8 GiB | 4-bit, smallest footprint |
The cache costs 120 KiB per token at 16-bit, so the full 262,144-token context of one request needs about 30.0 GiB. Concurrency multiplies that number, not the weights.
Hardware
At the released precision, with 90% of the card given to the server and 2.5 GiB kept for the runtime. The context column is what is left for the cache on that setup.
| GPU | Cards needed | Free for cache | Context it holds |
|---|---|---|---|
| RTX 4090 24GB | more than 8 | — | — |
| RTX 5090 32GB | more than 8 | — | — |
| L40S 48GB | more than 8 | — | — |
| A100 80GB | 8 | 176.9 GiB | the full 262,144 tokens |
| H100 80GB | 8 | 176.9 GiB | the full 262,144 tokens |
| RTX PRO 6000 96GB | 8 | 292.1 GiB | the full 262,144 tokens |
| H200 141GB | 4 | 118.5 GiB | the full 262,144 tokens |
Serving
A vLLM launch line for the shape above. Check the model card for a runtime the publisher recommends.
vllm serve nex-agi/Nex-N2.5-Pro \ --tensor-parallel-size 8 \ --max-model-len 262144 \ --gpu-memory-utilization 0.90
Questions
The weights take about 379 GiB at the released precision. No single card in the table above holds that, so it needs several GPUs or a lower precision.
The repository declares apache-2.0. Read the licence text before commercial use: the name alone does not tell you what is allowed.
The configuration allows 262,144 tokens. Whether the whole window is usable depends on the memory left for the cache, which the hardware table works out per card.
This page is a reference, and Aiqre does not serve this model today. We deploy open-weight models on dedicated EU hardware on request, so if you want an OpenAI-compatible endpoint for it, email [email protected] with the model name and roughly what volume you expect.
Also in the library
Nex-N2.5-mini is a 35.1 billion-parameter mixture-of-experts language model released under Apache-2.0.
Nex-N2.5-Max is a 1.6-trillion-parameter text-only MoE model designed for long-horizon agentic tasks.
GigaChat 3.5 Reasoning is a 438 billion-parameter mixture-of-experts language model trained with online RL for improved reasoning…
K2-Horizon-375B-A23B is an open-weight sparse MiE language model from IFM with 379 B parameters, 23 B active per token and a 512 …
Hosting
Tell us the model and roughly what volume you expect. We reply with a price and an OpenAI-compatible endpoint, with your prompts never stored.