| Revision as of 19:37, 21 December 2024 editVeldsenk (talk | contribs)Extended confirmed users2,171 editsNo edit summary← Previous edit | Revision as of 19:46, 21 December 2024 edit undoVeldsenk (talk | contribs)Extended confirmed users2,171 edits Filled in 1 bare reference(s) with reFill 2Next edit → | ||
| Line 1: | Line 1: | ||
| '''OpenAI o3''' is a ] model developed by ] as a successor to the ] model. It is designed to devote additional deliberation time when addressing questions that require step-by-step logical reasoning.<ref>https://www.wired.com/story/openai-o3-reasoning-model-google-gemini/</ref><ref>https://www.nytimes.com/2024/12/20/technology/openai-new-ai-math-science.html</ref> | '''OpenAI o3''' is a ] model developed by ] as a successor to the ] model. It is designed to devote additional deliberation time when addressing questions that require step-by-step logical reasoning.<ref name="auto">{{Cite web|url=https://www.wired.com/story/openai-o3-reasoning-model-google-gemini/|title=OpenAI Upgrades Its Smartest AI Model With Improved Reasoning Skills|first=Will|last=Knight|via=]}}</ref><ref>https://www.nytimes.com/2024/12/20/technology/openai-new-ai-math-science.html</ref> | ||
| ==History== | ==History== | ||
| OpenAI o3 model was announced on December 20, 2024, with the designation "o3" chosen to avoid trademark conflict with the existing UK mobile carrier named ].<ref |
OpenAI o3 model was announced on December 20, 2024, with the designation "o3" chosen to avoid trademark conflict with the existing UK mobile carrier named ].<ref name="auto"/> The model is available in two versions: o3 and o3-mini, and access is provided through an invitation-based testing program.<ref name="auto"/> | ||
| ==Capabilities== | ==Capabilities== | ||
| o3 demonstrates improved performance over the o1 model in complex tasks, including ], ], and ].<ref |
o3 demonstrates improved performance over the o1 model in complex tasks, including ], ], and ].<ref name="auto"/> On the ARC-AGI benchmark, which evaluates an AI's ability to handle new, challenging mathematical and logical problems, o3 attains three times the accuracy of its predecessor.<ref name="auto"/> | ||
| ==References== | ==References== | ||
Revision as of 19:46, 21 December 2024
OpenAI o3 is a generative pre-trained transformer model developed by OpenAI as a successor to the OpenAI o1 model. It is designed to devote additional deliberation time when addressing questions that require step-by-step logical reasoning.
History
OpenAI o3 model was announced on December 20, 2024, with the designation "o3" chosen to avoid trademark conflict with the existing UK mobile carrier named O2. The model is available in two versions: o3 and o3-mini, and access is provided through an invitation-based testing program.
Capabilities
o3 demonstrates improved performance over the o1 model in complex tasks, including coding, mathematics, and science. On the ARC-AGI benchmark, which evaluates an AI's ability to handle new, challenging mathematical and logical problems, o3 attains three times the accuracy of its predecessor.
References
- ^ Knight, Will. "OpenAI Upgrades Its Smartest AI Model With Improved Reasoning Skills" – via Wired.com.
- https://www.nytimes.com/2024/12/20/technology/openai-new-ai-math-science.html
| OpenAI | |||||||||
|---|---|---|---|---|---|---|---|---|---|
| Products | |||||||||
| Foundation models | |||||||||
| People |
| ||||||||
| Related | |||||||||
