AI model/language model | OpenAI goes open source: “Very, very good model” comes without restrictions – Image: Xpert.Digital
OpenAI plans an open-source model without restrictions
OpenAI announces new open-source language model
OpenAI plans to release a new, powerful, open-source language model with enhanced logical reasoning capabilities in the coming months. This announcement marks a significant strategic shift for the company, which has not released a fully open model since GPT-2. CEO Sam Altman describes the upcoming model as a “very, very good model” and emphasizes the intention to make it available without restrictive terms of use.
Related to this:
The return to open-source roots
After years of focusing on proprietary models like GPT-3 and GPT-4, this announcement represents a remarkable U-turn in OpenAI's strategy. With the release of GPT-3, the company had largely distanced itself from open source, citing security concerns. Now, CEO Sam Altman admits that this recent, rather closed-minded stance towards open source may have been problematic.
In a recent Reddit Q&A, Altman stated that he personally believes OpenAI needs to “develop a different open-source strategy,” even if it isn't currently the company's top priority. He added that while OpenAI will continue to produce better models, its lead will be smaller than in previous years.
This change of course comes at a time when competitive pressure on OpenAI is increasing from other AI developers such as DeepSeek, Meta, Alibaba, and Mistral, which have already successfully launched open models. In particular, the success of DeepSeek's R1 model and Meta's Llama family is seen as a potential catalyst for OpenAI's decision.
Related to this:
Developer involvement and community engagement
To make the new model as useful as possible, OpenAI is actively seeking dialogue with developers, researchers, and the wider community. The company has published a feedback form on its website, asking questions such as “What would you like to see in an OpenAI OpenWeight model?” and “What open models have you used in the past?”.
OpenAI is also planning a series of developer events, starting in San Francisco and later continuing in Europe and the Asia-Pacific region. These events will allow developers to test early versions of the model and provide direct feedback.
Brad Lightcap, COO of OpenAI, explains that the decision to adopt an open-source model is primarily based on strong demand from developers as well as business and government customers. “While we will continue to offer top-tier models through our API and in ChatGPT, there are many scenarios where APIs alone don't provide developers with the full flexibility they need,” says Lightcap.
Special features and terms of use
The new model is said to possess special capabilities in the area of logical reasoning, similar to OpenAI's o3-mini model. It is also noteworthy that Altman emphasizes the model will be released without restrictive terms of use.
In an apparent dig at Meta, he stated: “We will not introduce any silly restrictions, such as limiting the service to no more than 700 million monthly active users.” This limit corresponds exactly to the usage restriction of Meta’s Llama model. While Llama has been criticized for not being truly open source, since Meta, in addition to the license restrictions, also does not disclose the training data, it remains to be seen whether OpenAI will only publish the model weights or actually make all components available for a complete replication of the system, including the training data.
Safety aspects and evaluation
Before publication, the model will be evaluated according to the OpenAI Preparedness Framework. This evaluation will pay particular attention to the specific risks of an open-source model, such as potential hostile modifications by third parties.
Johannes Heidecke of OpenAI emphasizes that the company conducts security tests, both internally and with external experts. OpenAI will not publish any models that could pose catastrophic risks, Heidecke states. This additional diligence is necessary because it is known that the model could be modified by third parties after its publication.
Impact on the AI landscape
OpenAI's decision to release another open-source model could have far-reaching implications for the AI landscape. In an increasingly competitive industry, the winner of the open-source race could have the advantage of making proprietary offerings obsolete and becoming the dominant platform among developers, much like Google did with Android.
For developers, OpenAI's high-quality open-source model means greater flexibility for specific use cases such as specialized fine-tuning, adjustable latency, local installations, or applications requiring full control over the data. These scenarios are not currently optimally served by APIs alone.
OpenAI focuses on open source: A turning point in AI strategy?
OpenAI's announcement of a new open-source language model with reasoning capabilities marks a significant shift in the company's strategy. Faced with increasing competitive pressure and growing demand for open models, OpenAI appears to be reassessing its position in the AI market.
The coming months will show whether the new model can live up to the high expectations and how it compares to other open-source alternatives. By actively involving the developer community and foregoing restrictive terms of use, OpenAI could take an important step towards bringing its original mission – the democratization of artificial intelligence – back into sharper focus.
Related to this:
Your global marketing and business development partner
☑️ Our business language is English or German
☑️ NEW: Correspondence in your native language!
Konrad Wolfenstein
I and my team are happy to be available to you as your personal advisor.
You can contact me by filling out the contact form here wolfenstein@xpert.digital:or simply call me at +49 7348 4088 965. My email address is
I'm looking forward to our joint project.
