o3 Price Falls by 80% And o3-Pro Rolls Out Today

The AI world just got a little more interesting. On June 10, 2025, OpenAI delivered two major updates: a massive 80% price drop for its o3 model and the release of o3-Pro, a more reliable and intelligent version now available in ChatGPT Pro and API.

This update can reshape the way developers, educators, and small startups build with AI. If you’re an AI developer or someone running production-scale tasks, these changes are likely to reshape how – and how much – you use OpenAI’s reasoning models.

Let’s explore what actually changed, what it means for people using AI in the real world, and what to expect next.

Try ChatGPT Free with unlimited use without any registration or sign-up.

What’s o3, and Why People Are Paying Attention

The o3 model is OpenAI’s most advanced reasoning model to date. It’s not just good at chatting, it’s good at thinking through things.

While previous models were fast talkers, o3 is more like a quiet analyst. It reasons through complex questions, handles longer instructions better, and sticks closer to the facts. Since its launch, it’s been powering ChatGPT for Plus users.

This model has quietly become a workhorse for anyone building tools that need reliable, intelligent answers.

What is the New Price of O3?

OpenAI slashed the price of using o3 across the board:

– Input tokens now cost $2 per million,
– Output tokens dropped to $8 per million,
– And cached outputs (for repeated responses) now go as low as $0.50 per million.

Say you’re summarizing a 20-page research paper, writing blog post drafts, or analyzing survey data with o3 reasoning model. That workflow just became more affordable without needing to switch to a simpler model.

For context, that’s one-fifth the previous cost, making o3 one of the most cost-effective large language models with reasoning capability on the market today. If you’re running high-volume tasks like summarizing documents or generating code, the savings can be substantial.

OpenAI Released o3-Pro – Advanced Variant of o3

Alongside the price drop, OpenAI also rolled out o3-Pro – a variant of the o3 model designed to be more focused and better at handling complex or extended tasks.

While o3 is great for general use, the Pro model steps in when accuracy, clarity, and consistency matter most. As OpenAI explains:

“Like o1-pro, o3-pro is a version of our most intelligent model, o3, designed to think longer and provide the most reliable responses… especially in math, science, and coding.”

The Pro variant uses the same model weights as o3 but runs in a different configuration, giving it stronger performance in tasks that require sustained attention.

What’s Special About o3-Pro?

– Improved accuracy and instruction-following
– More comprehensive and clearer responses
– Outperforms o1-Pro and o3 in expert and academic evaluations
– Ideal for science, programming, business logic, and writing help

Unlike its faster siblings, the Pro model takes a little more time to respond because it’s built to reason through complex problems, not just reply quickly.

What AI Experts/Reviewers Say About the Pro Version?

Expert evaluations consistently preferred o3-Pro over o3 in every tested domain. Reviewers rated it higher for:

– Instruction adherence
– Scientific precision
– Business reasoning
– Writing support
– Code correctness

OpenAI’s own 4/4 reliability tests, where a model must answer a question correctly four times in a row, favor o3-Pro over both o3 and o1-Pro.

So, if you care more about reliability than speed, the Pro is the model to pick.

Where Can You Use o3-Pro?

ChatGPT Pro subscribers, at $200/month, and API users (via pay‑as‑you‑go or Pro-tier) have immediate access to o3‑Pro.

It’s also rolling out to ChatGPT Team, ChatGPT Enterprise, and Education accounts – Team and Enterprise users get access now, with Education following shortly.

If you’re a ChatGPT Plus subscriber ($20/month), you still get o3, but not the Pro version of the model. Think of o3-Pro as the premium tier for those who need higher precision – data analysts, researchers, engineers, or power users who can’t afford mistakes.

What is the Price of O3-Pro Reasoning Model?

o3-Pro Reasoning Model costs $20.00 per million input tokens and $80.00 per million output tokens. This means o3-Pro is 10x more expensive than standard o3. This pricing is specific to API usage, not the ChatGPT product tiers.

This pricing reflects this Pro model’s design for mission-critical tasks rather than everyday chat. It’s 10x costlier than its predecessor model, but the trade-off is performance and precision.

While both models use the same architecture, the Pro one runs in a more focused and deterministic configuration (sometimes referred to internally as “low-temperature” or “flex mode”), making it more suitable for production-grade outputs where precision and stability matter more than creativity or diversity.

Limitations of o3 Pro Model

OpenAI has been transparent about what o3-Pro can’t do just yet:

It can’t generate images – you can use GPT-4o or o3 for that. It doesn’t offer canvas support – visual interfaces and design tools are currently unsupported. Temporary chats are disabled in the Pro version due to a technical issue, which is being resolved by OpenAI.

This pro model produces slightly slower responses, especially when using tools like Python or file analysis.

Which Tools Pro Version of o3 Can Access?

Despite its limitations, o3-Pro comes fully loaded with all the advanced features of o3:

– File uploads and analysis
– Python for computation
– Web browsing
– Visual input interpretation
– Memory for personalization

This makes o3-Pro not just smarter, but also more versatile than o1-Pro and most other models.

Who Benefits Most from These Changes?

It’s easy to assume this only helps big companies running AI on a scale. But it’s actually the independent builders, teachers, creators, and researchers who stand to gain the most, especially when it comes to the use of the pro model.

– A solo developer building a productivity app can now run queries with o3 for a fraction of the former cost.
– A science teacher using AI to prep lesson plans or quizzes can do so more affordably.
– A data journalist generating summaries, timelines, or context for stories can handle heavier workloads with o3-Pro.

This kind of pricing unlocks workflows that would’ve been too expensive to automate last month.

What Happens Next?

This isn’t the end of the road; it’s more like a shift in direction. We might see AI tools and plugins built on o3 become cheaper to run, which could push subscription prices down for users.

The newly launched Pro reasoning model might quietly become the new standard for teams handling sensitive or technical work.

OpenAI could use this as a stepping stone to launch more “Pro-mode” variants, each tailored for different high-focus tasks.

This release also casts a long shadow over the rumored GPT-5. With o3-Pro now available and smarter memory features rolling out soon, OpenAI might not need a headline-grabbing new model just yet.

Final Thought

By dropping prices and offering a more stable version of its top model, OpenAI is signaling that access, not just power, is the next phase of AI’s evolution. And in doing so, they’re putting smart tools within reach for a much wider range of people.

If you’re a developer building applications where precision matters, or a researcher who can’t afford hallucinated responses, o3-Pro is now the most robust choice from OpenAI.

But if you’re writing blog posts, chatting casually, or just exploring AI for fun, the now-cheaper o3 model gives you plenty of power for a fraction of the price.