ChatGPT May Gain Video Power
OpenAI is reportedly preparing to fold Sora’s advanced video-generation tech into ChatGPT, a move that could let the model both produce and parse video for training or interfaces — potentially speeding embodied AI research reported. That capability would blur lines between vision models and LLMs and create new data pipelines for robotics simulation and policy learning.
The Information reported theinformation.com on March 10, 2026 that OpenAI is preparing to surface its Sora video-generation features directly inside ChatGPT. OpenAI published Sora 2 as its flagship video+audio model on Sept. 30, 2025. openai.com Appfigures data showed Sora’s installs fell 45% month‑over‑month to about 1.2 million in January 2026, according to TechCrunch. techcrunch.com The Information also reported OpenAI projects it could spend more than $225 billion on inference (model execution) from 2026–2030, a figure cited as a key cost pressure behind the product shift. theinformation.com OpenAI announced ChatGPT had surpassed 900 million weekly active users on Feb. 27, 2026, framing the integration as a route to scale Sora’s reach. openai.com The Information says OpenAI plans to keep the standalone Sora app while adding Sora features to ChatGPT, theinformation.com and PCMag notes current Sora access uses limited free daily generations with paid Pro/Plus/Business tiers for expanded video quotas — a credit-based monetization model that could be extended into ChatGPT. pcmag.com Reuters coverage cited by multiple outlets also flags Sora as a direct competitor to Meta and Google’s text‑to‑video efforts. money.usnews.com
