Nex-N2-Pro launches as a 397B MoE reasoning model based on Qwen3.5

Nex-N2-Pro 发布：基于 Qwen3.5 的 397B MoE 推理模型，性能达 GPT-5.5 水平

neolab released Nex-N2-Pro, a 397B-parameter MoE reasoning model based on Qwen3.5-397B-A17B, with 262K context, VLM support, claimed GPT-5.5 and Claude Opus 4.7-level performance, 30–50% fewer thinking tokens, SOTA results on Terminal Bench 2.1, GDPVal, and SWE-Verified, plus free access for the first two weeks via SiliconFlow.

Why it matters: HKR-H/K/R pass: the title has a strong benchmark hook and the post gives size, context, and token-reduction claims. Kept in 72-77 because it is a single X source and evaluation conditions are not disclosed.