AI Briefing
KOSign in

InternScience Releases Sage-A1.5, a 35B MoE Agentic Model for Scientific Discovery

·2026.10.21 17:59

Key point

Built on Qwen3.6-35B-A3B, the model features agentic orchestration and verification for long-horizon tasks.

Details

  • Sage-A1.5 is a 35B Mixture-of-Experts agentic model based on Qwen3.6-35B-A3B from Shanghai AI Laboratory.
  • The model is designed to advance long-horizon scientific problem solving through adaptive capability composition.
  • Key capabilities include Agentic Orchestration, Verification, and Open Scientific Discovery.
  • Agentic Orchestration enables task delegation and parallel collaboration among subagents with isolated contexts.
  • Verification Mode allows the model to pause mid-trajectory to review progress and revise strategy without an external critic.
  • The model natively supports function calling and tool integration for APIs, code interpreters, and search engines.
  • GGUF versions are available in Q4_K_M and Q8_0 quantizations.
open-sourcemixture-of-expertsagentic-aiscientific-discoverysage-a1.5qwen3.6-35b-a3b
Read originalShare on Threads

This summary was generated automatically by AI. Check the original for the author's claims and context. Copyright belongs to the original author.

Our guide explains how the AI works. Report summary errors, attribution issues, or removal requests via Contact.

Related

RedditAgnes AI Releases Agnes-3.0-Qwen: 180B MoE Model with 1M Context and Strong Agentic BenchmarksRedditinclusionAI Releases Ming-3.0-audio-tiny, a 9.3B Parameter Native Audio-Language ModelKakaoKakao Releases Kanana-2-vision-encoder, Open-Sourcing Medium and Tiny WeightsGeekNewsArtCraft releases open-source IDE for AI image and video creation with 62-model catalogDeepSeekDeepSeek Releases V4.1-Flash-1017 and Opens Search & Fetch APIsPyTorchKRStrata Engine Released to Run Qwen3.8-Flash-Next 125B on 12GB VRAM