OpenAI introduced Tuesday the launch of two open-weight AI reasoning fashions with comparable capabilities to its o-series. Each are freely obtainable to obtain from the web developer platform Hugging Facethe corporate mentioned, describing the fashions as “cutting-edge” when measured throughout a number of benchmarks for evaluating open fashions.
The fashions are available two sizes: a bigger and extra succesful gpt-oss-120b mannequin that may run on a single Nvidia GPU, and a lighter-weight gpt-oss-20b mannequin that may run on a shopper laptop computer with 16GB of reminiscence.
The launch marks OpenAI’s first ‘open’ language mannequin since GPT-2, which was launched greater than 5 years in the past.
In a briefing, OpenAI mentioned its open fashions might be able to sending advanced queries to AI fashions within the cloud, as TechCrunch beforehand reported. Which means if OpenAI’s open mannequin shouldn’t be able to a sure process, equivalent to processing a picture, builders can join the open mannequin to one of many firm’s extra succesful closed fashions.
Whereas OpenAI open sourced AI fashions in its early days, the corporate has typically favored a proprietary, closed supply growth strategy. The latter technique has helped OpenAI construct a big enterprise promoting entry to its AI fashions through an API to enterprises and builders.
Nonetheless, CEO Sam Altman mentioned in January he believes OpenAI has been “on the fallacious aspect of historical past” on the subject of open sourcing its applied sciences. The corporate as we speak faces rising stress from Chinese language AI labs — together with DeepSeek, Alibaba’s Qwen, and Moonshot AI — which have developed a number of of the world’s most succesful and common open fashions. (Whereas Meta beforehand dominated the open AI area, the corporate’s Llama AI fashions have fallen behind within the final 12 months.)
In July, the Trump administration additionally urged U.S. AI builders to open supply extra know-how to advertise world adoption of AI aligned with American values.
Techcrunch occasion
San Francisco
|
October 27-29, 2025
With the discharge of gpt-oss, OpenAI hopes to curry favor with builders and the Trump administration alike, each of which have watched the Chinese language AI labs rise to prominence within the open supply area.
“Going again to after we began in 2015, OpenAI’s mission is to make sure AGI that advantages all of humanity,” mentioned Altman in an announcement shared with TechCrunch. “To that finish, we’re excited for the world to be constructing on an open AI stack created in the USA, based mostly on democratic values, obtainable without spending a dime to all and for large profit.”
Picture Credit:Tomohiro Ohsumi / Getty Pictures
How the fashions carried out
OpenAI aimed to make its open mannequin a pacesetter amongst different open-weight AI fashions, and the corporate claims to have performed simply that.
On Codeforces (with instruments), a aggressive coding check, gpt-oss-120b and gpt-oss-20b rating 2622 and 2516, respectively, outperforming DeepSeek’s R1 whereas underperforming o3 and o4-mini.
OpenAI’s open mannequin efficiency on codeforces.Picture Credit:OpenAI
On Humanity’s Final Examination (HLE), a difficult check of crowdsourced questions throughout a wide range of topics (with instruments), gpt-oss-120b and gpt-oss-20b rating 19% and 17.3%, respectively. Equally, this underperforms o3 however outperforms main open fashions from DeepSeek and Qwen.
OpenAI’s open mannequin efficiency on HLE.Picture Credit:OpenAI
Notably, OpenAI’s open fashions hallucinate considerably greater than its newest AI reasoning fashions, o3 and o4-mini.
Hallucinations have been getting extra extreme in OpenAI’s newest AI reasoning fashions, and the corporate beforehand mentioned it doesn’t fairly perceive why. In a white paper, OpenAI says that is “anticipated, as smaller fashions have much less world data than bigger frontier fashions and have a tendency to hallucinate extra.”
OpenAI discovered that gpt-oss-120b and gpt-oss-20b hallucinated in response to 49% and 53%, respectively, of questions on PersonQA, the corporate’s in-house benchmark for measuring the accuracy of a mannequin’s data about folks. That’s greater than triple the hallucination charge of OpenAI’s o1 mannequin, which scored 16%, and better than its o4-mini mannequin, which scored 36%.
Coaching the brand new fashions
OpenAI says its open fashions had been educated with comparable processes to its proprietary fashions. The corporate says every open mannequin leverages mixture-of-experts (MoE) to faucet fewer parameters for any given query, making it run extra effectively. For gpt-oss-120b, which has 117 billion complete parameters, OpenAI says the mannequin solely prompts 5.1 billion parameters per token.
The corporate additionally says its open mannequin was educated utilizing high-compute reinforcement studying (RL) — a post-training course of to show AI fashions proper from fallacious in simulated environments utilizing massive clusters of Nvidia GPUs. This was additionally used to coach OpenAI’s o-series of fashions, and the open fashions have the same chain-of-thought course of by which they take further time and computational assets to work by their solutions.
On account of the post-training course of, OpenAI says its open AI fashions excel at powering AI brokers and are able to calling instruments equivalent to internet search or Python code execution as a part of its chain-of-thought course of. Nonetheless, OpenAI says its open fashions are text-only, that means they will be unable to course of or generate photos and audio like the corporate’s different fashions.
OpenAI is releasing gpt-oss-120b and gpt-oss-20b beneath the Apache 2.0 license, which is mostly thought of some of the permissive. This license will enable enterprises to monetize OpenAI’s open fashions with out having to pay or get hold of permission from the corporate.
Nonetheless, not like totally open supply choices from AI labs like AI2, OpenAI says it won’t be releasing the coaching information used to create its open fashions. This choice is no surprise on condition that a number of energetic lawsuits in opposition to AI mannequin suppliers, together with OpenAI, have alleged that these firms inappropriately educated their AI fashions on copyrighted works.
OpenAI delayed the discharge of its open fashions a number of occasions in latest months, partially to handle security considerations. Past the corporate’s typical security insurance policies, OpenAI says in a white paper that it additionally investigated whether or not dangerous actors may fine-tune its gpt-oss fashions to be extra useful in cyberattacks or the creation of organic or chemical weapons.
After testing from OpenAI and third-party evaluators, the corporate says gpt-oss could marginally improve organic capabilities. Nonetheless, it didn’t discover proof that these open fashions may attain its “excessive functionality” threshold for hazard in these domains, even after fine-tuning.
Whereas OpenAI’s mannequin seems to be state-of-the-art amongst open fashions, builders are eagerly awaiting the discharge of DeepSeek R2, its subsequent AI reasoning mannequin, in addition to a brand new open mannequin from Meta’s Superintelligence Lab.
