GPT-5.4 mini and nano followed GPT-5.2; this profile remains specific to GPT-5.2

The OnAirToday AI Progress Tracker recorded a verified change to OpenAI's GPT-5.2 (version gpt-5.2): GPT-5.4 mini and nano followed GPT-5.2; this profile remains specific to GPT-5.2.

GPT-5.2 is an OpenAI reasoning model supporting text and image inputs, long-context analysis, coding, and tool-driven workflows. Its published launch evaluations show improvements over GPT-5.1, although benchmark results depend on reasoning settings and do not guarantee reliable autonomous operation.

What changed in this review:

• GPQA Diamond and SWE-bench Verified now use OpenAI's GPT-5.2 Thinking launch results, not unattributed evaluation configurations.
• Added published AIME 2025 and ARC-AGI-2 launch results.
• Removed unsupported benchmark entries; differing evaluation configurations should not be interpreted as performance changes.
• Added later-family security reports without attributing their incidents to GPT-5.2.
• Prices are US dollars per million tokens for standard API input and output, excluding cached-input discounts.
• Flaws and limitations updated

Teams building on GPT-5.2 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/gpt-5-2.

References:

• https://openai.com/index/introducing-gpt-5-2/
• https://platform.openai.com/docs/models/gpt-5.2
• https://openai.com/api/pricing/
• https://openai.com/index/introducing-gpt-5-4-mini-and-nano