OpenAI has officially unveiled GPT-5, integrating advanced reasoning from models like o3 into a unified powerhouse now available through ChatGPT and its API.
The launch follows early hints and marks a pivotal shift toward streamlined AI capabilities.
Strategic PivotCEO Sam Altman teased GPT-5 as early as February 2025, positioning GPT-4.5 as the last non-chain-of-thought model.
ALSO READ Climate-resilient crops: Gates Foundation grants $7m to advance research in Africa, South-Asia
Summer rollout rumors solidified in July, with Altman posting screenshots in August that showcased agentic tasks and multimodal prowess.
Rollout timeline
The model launched August 6, 2025, quickly extending to all ChatGPT users. Platforms like Perplexity adopted variants such as GPT-5.2 with thinking features by December.
Major GPT-5 upgrades
GPT-5 fuses key breakthroughs for enhanced task performance, supported by 5,000 hours of safety training to reduce hallucinations. It handles vibe coding from simple prompts and transparently notes its limits.
