HeyGen's Decision 1 lands at 93 ms median
HeyGen says its new decision model answers at about 93 ms median on Turbo and leads on audio tasks like speech emotion, 84.5 against 32.4. Every number in that launch came from HeyGen, and the source never says what it was compared against.
Built with it
TypeSafe AI says Jev decides insurance risk for Strala Group
TypeSafe AI says Jev gives Strala Group the reliability they need to power decision-making for insurance, where quantifying risk is as critical as it gets. The evidence is an interview with Armando Strala and @flohop_dev, not a write-up: no accuracy, no latency, no cost, no description of what Jev replaced.
Someone measured it
HeyGen's Decision 1 goes live on OpenRouter at 93 ms median
Decision 1 from HeyGen is on OpenRouter, taking up to 400 questions about video, audio, images and text in one request and returning a calibrated probability for each.
Ask questions about video, audio, images, and text in one request, and get a calibrated probability for each answer in ~100 ms.
