Nex AGI releases Nex-N2.6, a 321B-parameter agentic model with 262K context at $0.12/$0.40 per 1M tokens

Nex AGI released Nex-N2.6 on Nov 10, 2026, an agentic model built on GLM-5.3-Flash with about 321B parameters and a 262K-token context window. OpenRouter lists it at $0.12 per 1M input tokens and $0.40 per 1M output tokens, labeled 20% off.

Key specifications

• Model: Nex-N2.6 (nex-agi/nex-n2.6)
• Developer: Nex AGI
• Base: GLM-5.3-Flash
• Parameters: about 321B
• Context window: 262K tokens
• Release date: Nov 10, 2026
• Pricing: $0.12 input / $0.40 output per 1M tokens (20% off label on OpenRouter)
• Cache read: $0.024 per 1M tokens listed
• Benchmark scores: none listed on the model page
• Training cutoff: not listed

What Nex AGI says it does

Nex AGI describes Nex-N2.6 as an agentic model suited for computer use, coding, and long-horizon tasks that involve tools. Compared with Nex-N2.5-Pro, the company says the new model reduces repetitive reasoning and follows the tools available in the current environment more closely. No benchmark figures accompany these claims.

The page links to model weights and includes a reasoning effort filter.

Pricing and availability

Nex AGI is the sole provider listed on OpenRouter. The listed rates are $0.12 per 1M input tokens and $0.40 per 1M output tokens, with cache reads at $0.024 per 1M tokens. The model is not currently available through EU in-region routing.

For comparison, the same page lists earlier Nex AGI models at these prices per 1M input / output tokens:

• Nex-N2.5-Pro: $0.075 / $0.25 (262K context)
• Nex-N2.5-Mini: $0.025 / $0.10 (262K context)

Live performance figures

As read on OpenRouter on Nov 10, 2026, the Nex AGI endpoint showed P50 latency of 2.95s and throughput of 30 tokens per second. OpenRouter availability over the preceding 24 hours was 90.47%. Per-provider uptime was not shown. These figures reflect launch-day conditions and will change.

Open WebUI was the top app using the model at the time of reading, with 8K tokens.

Lineage

Nex-N2.6 follows the Nex-N2.5 releases on OpenRouter. The earlier Nex-N2-Pro was a mixture-of-experts model with 17B active parameters out of 397B total, built on the Qwen3.5 architecture. Nex-N2.6 uses a different base, GLM-5.3-Flash, and Nex AGI gives its size as about 321B parameters.

What this means

Nex-N2.6 targets agent workloads where the model must work within whatever tools a given environment exposes. Nex AGI's stated improvements, less repetitive reasoning and closer adherence to available tools, address behaviors that affect cost and reliability in long tool-using runs. Those claims are unverified until independent benchmarks or evaluations appear.

On price, Nex-N2.6 is listed above the Nex-N2.5-Pro and Nex-N2.5-Mini rates on the same page, so teams moving from those models should test whether the behavior changes justify the cost. The 262K context window matches the earlier Nex-N2.5 models.

Two practical constraints apply at launch: no EU in-region routing, and a single provider, which limits OpenRouter's ability to reroute around upstream errors. The 90.47% 24-hour availability reading is relevant for anyone planning production use.