💬 Try Nex-N2-Pro
Chat with the model powered by OpenRouter (free tier available)
Try these:
📊 Benchmarks
Nex-N2-Pro competes with top-tier proprietary models
| Benchmark | Nex-N2-Pro | GPT-5.5 | Opus 4.7 | DeepSeek-V4-Pro |
|---|---|---|---|---|
| BrowseComp | 83.7 | 84.4 | 79.8 | 83.4 |
| GDPval | 1585 | 1769 | 1753 | 1554 |
| Toolathlon | 51.9 | 55.6 | 52.8 | 51.8 |
| WildClawBench | 53.5 | 58.2 | 62.2 | 43.7 |
| WideSearch | 75.6 | - | - | - |
| TAU3 | 71.1 | - | - | - |
🧠 About Nex-N2-Pro
Adaptive Thinking
Decides on its own when to think and how deeply — executing simple actions quickly while reasoning thoroughly on critical decisions.
Coherent Thinking
Carries one consistent reasoning paradigm across general reasoning and diverse agentic tasks, enabling stable capability transfer.
Agentic Capabilities
First-tier coding, tool calling, deep research, and terminal execution — all unified in a single closed loop.
Open Source
Released under Apache 2.0. Available on Hugging Face and ModelScope. Two variants: Pro (397B) and mini (35B).