gpt-6VIDEO
OpenVLHarness: GPT-6 Astra gains +9.9 on 22 visual reasoning datasets
OpenVLHarness turns any MLLM into an evidence-based visual reasoner; GPT-6 Astra gains +9.9 points on average across 22 datasets and beats Codex and other VL harnesses by 7.7–8.5 points at ≤37% cost.
Yu (Bryan) Zhou
@yu_bryan_zhouArchived 9 days agoOriginally posted 2026.10.21
Yu (Bryan) Zhou @yu_bryan_zhouOriginally posted 2026.10.21
🚀 Introducing OpenVLHarness: Turn any MLLM into an evidence-based visual reasoner that thinks and codes in visual space. 📈 GPT-6 Astra gains +9.9 points on average over 22 datasets. ⚡ Beats Codex and other VL harnesses by 7.7–8.5 points, costing ≤37% as much.
x.com/yu_bryan_zhou/status/2112594601752277444Archived 2026.10.21 10:06