Which AI writes the better take? You decide — blind.

Glimpse exposes API access to a vision model that can re-inspect visual inputs during reasoning by cropping, pulling video frames, and running measurement code, priced at $0.60/M input tokens and $3/M output tokens. For teams building on image or video data, the practical shift is that some tasks currently handled with custom CV pipelines or human review—chart interpretation, short video captioning, board-state understanding—may become feasible through a general model, but at a cost tier above lightweight multimodal models.

A purpose-built classification model that picks among predetermined outputs hit 35% of Vercel's paid teams in two weeks, outpacing Sonnet 5.5 roughly 2-to-1 — Vercel itself swapped it in for GPT-6 Luna on harmful-command filtering. If you're running general-purpose frontier models for routing, moderation, or bucketing high-volume data, you're overpaying and underperforming; constrained-output models like this are now the default-cheaper, faster option for that tier of work, with OpenAI and Amazon already shipping competitors.

China is reportedly planning to implement restrictions on the export of open-weights artificial intelligence models, limiting access to the underlying parameters that allow these systems to function. The move signals a potential shift in how the country manages the international distribution of its advanced AI technologies amid growing global competition in the sector.