ClosedOpenAI/10 trustVerified

o4-mini

Name: o4-mini
Rating: 9

The reasoning model that thinks with images, at o3-mini prices

Visit official page →

About

o4-mini arrived April 16, 2025 alongside o3 and immediately replaced o3-mini across OpenAI's lineup. The headline feature is thinking with images: rather than just describing a picture, the model pulls images into its chain of thought, rotating and manipulating them as it reasons. Combined with full agentic tool access (web search, Python, file analysis, image generation) and elite math scores at $1.10/$4.40 per million tokens, it became the default production reasoning model for cost-sensitive teams. Free ChatGPT users got it through a Think toggle a week later. Its main limitation is breadth: it is a specialist, not a generalist, and complex open-ended tasks still favor the larger models.

Specs

Context window

200,000 tokens

Pricing

$1.10 per 1M input tokens, $4.40 per 1M output tokens

Free quota

Free ChatGPT users via the Think toggle (from April 24, 2025)

API access

OpenAI API, Azure OpenAI

Weights

Closed — API only

Released

April 16, 2025

Benchmarks

Team-reported: AIME 2024 93.4%, AIME 2025 92.7%, SWE-bench Verified 68.1%

Provider-reported unless a source says otherwise. Benchmarks are marketing-adjacent by nature — treat them as directional.

Use cases

Visual reasoning over diagrams and sketchesMath-heavy production workloadsAgentic coding at scaleCost-efficient tool-using agents

Trust score — 9/10

/10 trustVerified

Scored 9/10 (verified tier, metrics-v1). Evidence: April 16, 2025 launch alongside o3 corroborated across 4+ sources; Thinking-with-images capability verified from multiple sources; AIME 93.4%/92.7% and SWE-bench 68.1% labeled team-reported; Free-tier Think toggle (Apr 24, 2025) captured; specialist-not-generalist caveat stated.