CaryptosHeadlines Media Has Launched Its Native Token CHT.
Airdrop Is Live For Everyone, Claim Instant 5000 CHT Tokens Worth Of $50 USDT.
Join the Airdrop at the official website,
CryptosHeadlinesToken.com
OpenAI has announced the launch of a new open-weight language model later this year, allowing developers to run the model on their own hardware. The company’s CEO Sam Altman shared the update on March 31, revealing that the company has been considering this step for a long time but had prioritized other projects until now.
The upcoming model will not be entirely open-source but will offer more flexibility compared to GPT-3 and GPT-4, which remain fully closed. It marks OpenAI’s first open-weight model since GPT-2 in 2019, signaling a shift towards greater accessibility.
Altman emphasized that the company is still making key decisions about the release and is hosting developer events in San Francisco, followed by sessions in Europe and the Asia-Pacific region, to gather feedback. The goal is to ensure the model is as useful as possible while considering potential risks associated with modifications.
OpenAI Competing in the AI Arms Race
The AI industry is seeing increased competition, with rivals like DeepSeek, which operates similarly to ChatGPT but was reportedly built at a fraction of the cost and time.
Meanwhile, Alibaba launched an open-source AI model on March 26, designed for cost-effective AI agents. Just a day earlier, Google introduced Gemini 2.5, the latest version of its experimental AI model.
OpenAI is also actively working on GPT-4.5 and GPT-5, with Altman previously hinting that they could launch in the coming weeks or months.
What This Means for AI Developers
An open-weight model offers developers greater control and customization opportunities compared to closed models. Companies and governments that prefer self-hosted AI solutions will now have an option from OpenAI rather than relying on proprietary cloud-based solutions.
Meta CEO Mark Zuckerberg also made headlines this month, revealing that Llama, Meta’s AI model family, has hit 1 billion downloads since its February 2023 release.
