o3: gPQA Diamond now reflects OpenAI's published 83.3% result, rather than the previously listed 81.8%

The OnAirToday AI Progress Tracker recorded a verified change to OpenAI's o3 (version o3 (o3-2025-04-16)): GPQA Diamond now reflects OpenAI's published 83.3% result, rather than the previously listed 81.8%.

OpenAI o3 is a reasoning model for complex coding, mathematics, scientific analysis and visual problem-solving, with text and image inputs and text outputs. Its documented API configuration provides a 200,000-token context window, with standard pricing of $2 per million input tokens and $8 per million output tokens.

What changed in this review:

• SWE-bench Verified now reflects OpenAI's published 69.1% result; differently configured evaluations are not directly interchangeable.
• Added OpenAI's published AIME 2024, AIME 2025 and MMMU results.
• OpenAI released o3-pro on June 10, 2025, as a separate higher-compute variant; this profile remains about o3.
• Flaws and limitations updated
• Risks and threats updated

Teams building on o3 should confirm limits, pricing and benchmark coverage against the vendor's own documentation before relying on these figures.

The full profile — capabilities, benchmark citations, known flaws and risk guidance — is on the tracker at /ai-models/o3.

References:

• https://platform.openai.com/docs/models/o3
• https://openai.com/index/introducing-o3-and-o4-mini/
• https://openai.com/index/o3-o4-mini-system-card/
• https://platform.openai.com/docs/models/o3-pro