The open-weight Video-DeepResearch 35B model has achieved a score of 64% on a new video reasoning benchmark, surpassing GPT-5's score of 52.5%. This model also outperformed Claude 4.5 Sonnet, though the research team graded their own benchmark results. AI
IMPACT This benchmark result suggests potential advancements in AI's video reasoning capabilities, challenging existing top models.
RANK_REASON The cluster reports on a new benchmark result for an open-weight model, which is a research milestone. [lever_c_demoted from research: ic=1 ai=1.0]
Read on Mastodon — fosstodon.org →
AI-generated summary · Google Gemini · from 1 sources. How we write summaries →
