Meet Nex-N2.6, our latest multimodal agentic model supporting text,...
Nex@NexEcosystem
7 views
Nov 11, 2026
~1 min read
Advertisement
1
Meet Nex-N2.6, our latest multimodal agentic model supporting text, image, and video inputs.
Open-sourced by Shanghai Innovation Institute, the model has performed strong benchmark results:
⚡Compared with Nex-N2.5-Pro, the reasoning repetition rate falls from 16.0% to 1.6%, while the overall non-repetition rate rises from 91.0% to 97.8%
⚡Achieved 100% valid tool selection rate across 956 tested contexts, on par with GPT-6-Sol
⚡Scored 83.0 on OSWorld and 84.1 on Terminal-Bench 2.1, ahead of DeepSeek-V4.1-Flash at 71.7 and 82.7
Nex-N2.6 is built with a stronger foundation to navigate interfaces, use tools, and complete lengthy tasks without losing its place.
Try now 👇
⭐GitHub: github.com/sii-nex/Nex-N2…
🤗HuggingFace: huggingface.co/nex-agi/Nex-N2…
🔗OpenRouter API: openrouter.ai/nex-agi/nex-n2…
🔗MOSS API: platform.moss.cn/docs/models/ne…
Open-sourced by Shanghai Innovation Institute, the model has performed strong benchmark results:
⚡Compared with Nex-N2.5-Pro, the reasoning repetition rate falls from 16.0% to 1.6%, while the overall non-repetition rate rises from 91.0% to 97.8%
⚡Achieved 100% valid tool selection rate across 956 tested contexts, on par with GPT-6-Sol
⚡Scored 83.0 on OSWorld and 84.1 on Terminal-Bench 2.1, ahead of DeepSeek-V4.1-Flash at 71.7 and 82.7
Nex-N2.6 is built with a stronger foundation to navigate interfaces, use tools, and complete lengthy tasks without losing its place.
Try now 👇
⭐GitHub: github.com/sii-nex/Nex-N2…
🤗HuggingFace: huggingface.co/nex-agi/Nex-N2…
🔗OpenRouter API: openrouter.ai/nex-agi/nex-n2…
🔗MOSS API: platform.moss.cn/docs/models/ne…
2
Nex-N2.6 is now live on OpenRouter
openrouter.ai/nex-agi/nex-n2…
openrouter.ai/nex-agi/nex-n2…
3
Harness cross-framework adaptation
Built to use the tools each framework actually provides.
🛠️ Across 956 tested contexts, Nex-N2.6 reached 100% valid tool selection rate
🔒 100% of its initiated tool calls avoided tools outside the available list
From Claude Code to Codex, Cursor, and other harnesses, the goal is straightforward: switch environments without losing track of which tools are available.
Built to use the tools each framework actually provides.
🛠️ Across 956 tested contexts, Nex-N2.6 reached 100% valid tool selection rate
🔒 100% of its initiated tool calls avoided tools outside the available list
From Claude Code to Codex, Cursor, and other harnesses, the goal is straightforward: switch environments without losing track of which tools are available.
4
Stronger computer use
Built for agents that need to operate a computer, not just describe the next step.
🖥️ Nex-N2.6 scores 83.0 on OSWorld, up from 82.2 for Nex-N2.5-Pro and ahead of GLM-5.3-Flash at 78.0
⌨️ It scores 84.1 on Terminal-Bench 2.1, up from 82.7 for Nex-N2.5-Pro
Stronger performance across desktop interfaces and terminal tasks.
Built for agents that need to operate a computer, not just describe the next step.
🖥️ Nex-N2.6 scores 83.0 on OSWorld, up from 82.2 for Nex-N2.5-Pro and ahead of GLM-5.3-Flash at 78.0
⌨️ It scores 84.1 on Terminal-Bench 2.1, up from 82.7 for Nex-N2.5-Pro
Stronger performance across desktop interfaces and terminal tasks.
