Models#Benchmark#Device control

China’s 1.6T open-source model Nex-N2.5 tops BrowseComp, beating GLM-5.3 and DeepSeek-V4 on agent work

Shanghai Innovation Institute released the Nex-N2.5 family on September 9, 2026 — 35B, 397B and 1.6T models, all under Apache-2.0. On the official table the 1.6T Max tops BrowseComp at 92.6, ahead of Claude Opus 5 and GPT-5.6 Sol, and outscores GLM-5.3 and DeepSeek-V4-Pro on automation and tool-use benchmarks.

Official benchmark chart: bar comparisons across six text benchmarks, with the three Nex-N2.5 sizes alongside Claude, GPT-5.6 Sol, GLM-5.3, Kimi-K3 and DeepSeek-V4-Pro

Shanghai Innovation Institute announced its Nex-N2.5 family as an open-source release on September 9, 2026, spanning a 35B Mini, a 397B Pro and a 1.6T Max, all under Apache-2.0. Max is post-trained on DeepSeek-V4-Pro-Base, and the team calls it its first complete post-training effort at trillion-parameter scale. On the comparison table published with the release it tops BrowseComp at 92.6 — the best score in the table, ahead of Kimi-K3 at 91.2, Claude Opus 5 at 90.8 and GPT-5.6 Sol at 90.4.

The facts

  • The lineup: Mini builds on Qwen3.5-35B-A3B and Pro on Qwen3.5-397B-A17B; both are multimodal MoE models aimed at computer use, web browsing and vision-driven self-correction. Max is a 1.6T-parameter, text-only MoE with 49B active, built on DeepSeek-V4-Pro-Base and aimed at complex reasoning, coding and agent work. All three have a 256K context window.
  • Background: Nex-AGI is an open-source ecosystem for agentic models that Shanghai Innovation Institute built with partner research groups and startups; Nex-N2.5 follows the Nex-N2 generation.
  • License and access: weights are on Hugging Face and ModelScope under Apache-2.0. Mini and Pro are also hosted on OpenRouter.
  • Benchmarks: the comparison table from the official README and site (updated September 8, 2026; bold marks the best score, — means unavailable). Text first:
BenchmarkMiniProMaxClaude Opus 5GPT-5.6 SolKimi-K3GLM-5.3DeepSeek-V4-Pro-0813Qwen3.8-Max
Terminal-Bench 2.173.482.786.189.188.888.388.287.986.6
SWE-Bench Pro43.861.265.779.264.663.364.655.467.7
DeepSWE v1.136.155.865.673.772.767.566.962.869.3
AutomationBench v1.0.632.344.250.250.345.846.748.243.239.8
Toolathlon Verified54.668.574.776.574.976.573.074.172.5
GDPval-AA v2144616281713183117111675176315801717
Job Bench28.541.453.665.745.452.958.254.153.4
BrowseComp83.489.792.690.890.491.2———

And multimodal:

BenchmarkMiniProMiniMax-M3Claude Opus 5GPT-5.6 SolKimi-K3GLM-5.3-FlashDeepSeek-V4-Flash-VisionQwen3.8-Max
OSWorld-Verified71.282.275.283.483.284.862.376.786.1
OSWorld-230.556.422.368.362.758.3——46.7
WebTest48.652.8——54.0———52.3
WebArena-Verified63.467.6——69.771.6—62.366.8
OSWorld-G82.987.4—76.877.779.683.359.484.9
Vision2Web52.968.259.0—79.8———75.1
SWE-MM25.538.2—59.440.237.320.639.239.2
OmniDoc89.792.291.6—92.991.1——92.1
  • Head to head: on the text table, Max scores higher than both GLM-5.3 and DeepSeek-V4-Pro-0813 on AutomationBench (50.2) and Toolathlon (74.7). It beats DeepSeek but trades wins with GLM on SWE-Bench Pro (65.7), DeepSWE (65.6) and GDPval (1713), and trails both on Terminal-Bench 2.1 (86.1) and Job Bench (53.6). On the multimodal table, Pro’s 87.4 on OSWorld-G is the top score.
  • Serving costs: the official Docker image (nexagi/sglang:v0.5.18-nex-patch) runs Mini on 2×H100, Pro on 8×H100, and Max on two nodes of 16×H200 each. Recommended sampling is temperature 0.7, top_p 0.95, top_k 40; reasoning_effort switches between no thinking, adaptive thinking and always thinking.
  • Scoring caveats: scores that have a public source come from model providers’ own reports (Kimi-K3, Qwen3.8-Max, GLM-5.3, HY4); the rest are Nex-AGI’s own runs. Computer-use benchmarks use its NexCUA harness, which is not open-sourced yet.
  • Demos: the team shows Max repairing a broken CRC16-CCITT streaming-accelerator RTL and taking it through synthesis and Sky130HD place-and-route (reported Fmax around 386 MHz, zero setup/hold TNS). Pro earns the first gym badge in Pokémon Platinum in 468-plus actions and turns a concept sketch into a 58-part mecha mesh in Blender.

Our take

On agent-style work this is the first Chinese open-weights release to beat every closed frontier model on a hard benchmark — at least on the official table, where BrowseComp 92.6 and OSWorld-G 87.4 are both the top scores including closed models. Read the numbers with care: most are run by the vendor, some competitor scores come from those competitors’ own reports, and none of it is third-party reproduced yet.

Direction by direction it looks narrower. On browser research, computer use and automation, Nex-N2.5 does beat GLM-5.3 and DeepSeek-V4-Pro-0813. On classic coding it trades wins — SWE-Bench Pro at 65.7 is ahead of both, but Terminal-Bench 2.1 and Job Bench lag behind, and Claude Opus 5’s 79.2 on SWE-Bench Pro is still far off.

The practical gate is hardware: Max needs 32 H200s, out of reach for most teams. What you can actually try is Mini or Pro on OpenRouter, or Mini on two H100s — a cheap way to test whether an open model can replace a closed API for computer-use work.

How to try it

  • Call nex-agi/nex-n2.5-pro or nex-agi/nex-n2.5-mini directly on OpenRouter
  • Download weights from Hugging Face or ModelScope (nex-agi/Nex-N2.5-Max, nex-agi/Nex-N2.5-Pro, nex-agi/Nex-N2.5-mini, all Apache-2.0)
  • Self-host with the official Docker image nexagi/sglang:v0.5.18-nex-patch; launch commands are in the README