US Edition
OpenAI launches new GPT-5 model for all ChatGPT users
- OpenAI unveiled GPT-5 on August 7, 2025, introducing it to ChatGPT and its API for developers and businesses worldwide.
- This launch follows advancements from previous models like GPT-4, with a unified architecture and improved reliability that reduces hallucinations and factual errors.
- GPT-5 offers enhanced capabilities including agentic coding, complex reasoning, multi-step tool use, free-form function calling, and a context window of 272,000 tokens.
- OpenAI claims GPT-5 reduces sycophantic behavior from 14.5% to under 6%, decreases the deception rate in conversations to 2.1%, and starts pricing at $1.25 per million input tokens.
- Although GPT-5 marks a significant step toward artificial general intelligence, OpenAI emphasizes it is not AGI or artificial superintelligence, with rollout available to free and paid subscribers and enterprises.
Insights by Ground AI
Does this summary seem wrong?
568 Articles
568 Articles
ChatGPT-5 has a higher reasoning capability than all previous versions submitted by OpenAI
On Thursday evening, the California-based AI company OpenAI released the latest version of ChatGPT. OpenAI CEO Sam Altman compared the launch of the new artificial intelligence with the invention of the iPhone. Does it promise too much? The new version of ChatGPT is here: GPT-5 is supposed to be more user-friendly and provide more accurate answers. Furthermore, the AI is supposed to answer with less "hallucinations" – i.e. false or freely invent…
·Vienna, Austria
Read Full ArticleCoverage Details
Total News Sources568
Leaning Left54Leaning Right41Center142Last UpdatedBias Distribution60% Center
Bias Distribution
- 60% of the sources are Center
60% Center
L 23%
C 60%
R 17%
Factuality
To view factuality data please Upgrade to Premium
