FlashNEWS BRIEFING

Shanghai AI Lab Open-Sources Scientific Multimodal Large Model InternLM-S2.1 Flash

1 min read

88Z AI, 17 November — Shanghai AI Laboratory has open-sourced the InternLM-S2.1 Flash scientific multimodal large model. Built on a unified representation architecture, it supports various scientific data inputs, combines scientific understanding with agent capabilities to directly control research software, significantly reduces inference costs, and introduces a recursive self-improvement mechanism, achieving excellent results in multiple general agent benchmarks.

At a glance

  • Supports various scientific data inputs and controls research software, with inference costs at one-tenth of InternLM-Xinghe
  • Introduces recursive self-improvement mechanism, ranking first in 4 and second in 7 out of 15 general agent benchmarks
Cover image for Shanghai AI Lab Open-Sources Scientific Multimodal Large Model InternLM-S2.1 Flash
Cover image for Shanghai AI Lab Open-Sources Scientific Multimodal Large Model InternLM-S2.1 Flash

The update

The Shanghai Artificial Intelligence Laboratory has officially open-sourced the InternLM-S2.1 Flash scientific multimodal large model. Built on the InternLM-Yuanzheng unified representation architecture, the model can process various scientific data inputs, including molecules, spectra, and biological sequences. It combines deep scientific understanding with robust agent capabilities, allowing it to directly control various research software to assist in scientific work. In terms of efficiency, the inference cost for the same task is only one-tenth that of the InternLM-Xinghe model, significantly lowering the barrier to use. Furthermore, the model introduces a recursive self-improvement mechanism, possessing self-evolution capabilities. It achieved excellent results in 15 general agent benchmarks, securing first place in four categories and second place in seven, demonstrating its leading strength in the field of scientific intelligence.

Sources & editorial note

This article is an AI-assisted synthesis by 88Z AI based on the public sources below. Product claims and availability are attributed to their publishers. It does not represent independent testing or first-hand reporting.

  1. 书生SageSource published

Source dates above refer to the original announcements; the date under the headline is this article’s publication on 88Z AI.

Related reading

All stories
88Z.ai