OpenAI showed Sora, a video model capable of long scenes with surprising visual coherence—still with artifacts.
In February 2024, OpenAI presented Sora demos, a diffusion/transformer video model from text. Up-to-a-minute clips showed camera coherence and apparent dynamics.
Hollywood, advertising, and education weighed opportunities and audiovisual deepfake risks. Initial access was limited.
Sora pushed competitors (Runway, Luma, Google) and asked whether video generators learn useful “world models” beyond pixels.
OpenAI launched GPT-4o with low-latency voice and native vision, pushing multimodal conversational UI.
The o1 series introduced inference with internal “chain of thought” optimized by RL, excelling at math and code.
OpenAI launched GPT-4 with notable gains on professional exams, robustness, and later visual ability.