OneSearch-VL: unified multimodal deep research agent beats Qwen3-VL by 20 points
Hongyu Li · hf · 2026-10-09
OneSearch-VL is a unified multimodal deep research agent built on a Visually Grounded Evidence Graph (VGEG) covering data construction, process supervision, and evaluation. With SFT-110K/RL-10K datasets and an Evidence-aware Visual-Grounded Rubric reward, OneSearch-VL-8B outperforms tool-augmented Qwen3-VL-8B by 20.2 and 17.6 points on two new benchmarks, plus gains across 7 image benchmarks and VideoDR.
More from coding & agent
- BrickBench benchmarks agentic LEGO design: agents pass constraints, trail humans — Peter Kulits · 2026-10-09
- autoicd-mcp ships automated ICD-10 medical coding MCP server with 74,000+ code search — modelcontextprotocol · 2026-10-09
- pohjola-api: agent-native Finnish company data API priced at $0.01 per call via x402 — modelcontextprotocol · 2026-10-09
- Dev recreates NES classic Faxanadu with Claude Opus, sharing daily progress — zeeg · 2026-10-09
- Every model looked bad in my eval — the bug was my answer key, not the models — jgarg27 · 2026-10-09
- Pi has no official Subagents, but 4 community extensions emerged; author shares extension stack order — solyarisoftware · 2026-10-09