Today for AI
HOT RADAR
agentHEAT 9.3°

AI CURATED HOT STORY · 11/14/2026

Alibaba Neos Open-Sources 9B Multimodal Search Agent Matching Closed Models

1 reports archived1 independent sourcesupdated 11/14/2026, 11:30:00
Synthesis & Latest Updates
1 Sources Cross-Validated

Alibaba's Accio team, with Fudan University and Shanghai Innovation Institute, introduces Neos, a unified training framework that couples data construction, process rewards and environment co-evolution. Its 9B and 35B models approach frontier closed-source performance on four multimodal search benchmarks by building multi-hop tasks with visual checkpoints and filtering out shortcut solutions.

LATEST/Alibaba's Accio team, with Fudan University and Shanghai Innovation Institute, introduces Neos, a unified training framework that couples data construction, process rewards and environment co-evolution. Its 9B and 35B models approach frontier closed-source performance on four multimodal search benchmarks by building multi-hop tasks with visual checkpoints and filtering out shortcut solutions.

Hourly heat curve

12 fully observed hours
Now
9.3
Peak
11.912:00
24h change
–

The line compares only sources observed throughout; gap hours are interpolated to keep the trend continuous.

Coverage timeline

Total 1 reports · Latest first
  1. 量子位 (微信公众号)T2·74 pts

    Alibaba Neos Open-Sources 9B Multimodal Search Agent Matching Closed Models

    • Neos unifies data construction, process rewards and environment co-evolution, letting a 9B model approach frontier closed-source search performance
    • Multi-hop tasks built from knowledge graphs embed visual entry points and intermediate checkpoints, with four-stage filtering to remove text-only or single-image shortcuts
    • Failed trajectories also feed credit assignment, so one expert trajectory supervises both decision-making and evidence handling