Intern-Decision: 0.8B-4B multimodal decision models output structured choices in one forward pass
internlm · hf · 2026-10-16
Shanghai AI Lab with SJTU, Fudan and NTU present Intern-Decision, a family of 0.8B/2B/4B multimodal decision models that return structured choices with candidate probabilities from text and optional images. Its autoregressive masked language modeling objective places one decision placeholder per field in a JSON skeleton, evaluating all fields in a single causal forward pass with the backbone's native LM head. The 4B model reaches 90.02% unweighted average accuracy across seven suites vs 88.74% for Jev v1.13.0, at much lower latency for a fixed 289-token request.
More from coding & agent
- SpecAgent: small model drafts tool calls, large model verifies in parallel to cut latency — Zilin Wang · 2026-10-16
- Salesforce's CLIFT turns sparse judge feedback into rewards for open-source web agents — Salesforce · 2026-10-16
- Agent spots invalid SPF record and autonomously fixes it in Vercel DNS via browser — dkundel · 2026-10-16
- MCPJam engineer asks how to test if ChatGPT plugins actually see latest UI selection — Paoli99 · 2026-10-16
- ChatGPT revives unplayable 2000s game Bookworm on Mac and iPhone in 15 minutes — Shleemy_Pants · 2026-10-16
- Agentflow: open-source collective memory network for agents via FastMCP and pgvector — agentflow0328 · 2026-10-16