Skip to content

supervisor: monitor InternScience HF org (Shanghai AI Lab science sub-org, Agents-A1) #38

Description

@SovereignSignal

Raised by: supervisor audit 2026-07-13

What

InternScience is the open-science sub-org of Shanghai AI Laboratory on HuggingFace (distinct from the main internlm org added in d798d0a). They open-sourced Agents-A1 on June 26, 2026 — a 35B MoE agentic model that achieves SotA on several agent benchmarks (Seal-0 56.4, HiPhO 46.4, FrontierScience-Olympiad 79.0) and competes with GPT-5.5 / DeepSeek-V4-Pro in the ~35B class.

Why not auto-committed

The authority caps require "sustained real model releases + meaningful engagement." Agents-A1 is InternScience's first major release under this org name (the team's sustained history is under internlm). One release does not meet the "sustained" bar on its own.

Recommended action

If InternScience releases a second major model (Agents-A2, Intern-S3, etc.), add them to KNOWN_ORGS, MAJOR_HF_ORGS, PROVIDER_NAMES, and significant_families in one focused commit. The intern-s pattern added to significant_families in d798d0a already covers their Intern-S series via trending, so the gap is modest.

References

Metadata

Metadata

Assignees

No one assigned

    Projects

    No projects

    Milestone

    No milestone

    Relationships

    None yet

    Development

    No branches or pull requests

    Issue actions