Physical Address
304 North Cardinal St.
Dorchester Center, MA 02124
Join the event, which is trustworthy for almost two decades. VB Transform brings together people who build a real AI strategy. Learn more
Good news, AI developer!
Openai has announced A Significant price reduction at O3its flagship large voice model (LMM), The cost of input and output tokens reduce the costs by a whopping 80%.
(Recall -token are the individual numerical consequences that LLMS use to present words, phrases, mathematical and coding strings and other content. They are representations of the semantic constructions that the model learned through the training. Essentially, the mother tongue of LLMS is essentially calculated for the privilege per million token).
The update positions the model as a better accessible option for developers who are looking for expanded argumentation functions and places Openai in direct prices with competing models such as Gemini 2.5 Pro from Google Deepmind, Claude Opus 4 from Anthropic and Deepseek’s argumentation suite.
Announced by Altman himself on X
Sam Altman, CEO of Openaai, confirmed the change in A Post on X Marked that the new pricing should promote a broader experimentation and wrote: “We have reduced the price of O3 by 80%!!
The costs for the use of O3 are now $ 2 per million in the input token and 8 USD per million output tokens with an additional discount of $ 0.50 per million token if the user enters information that has been “stored” or saved and identical to what he has previously made available and identical.
This marks a significant reduction compared to the previous interest rates of $ 10 (input) and 40 US dollars (issue) as Openai researcher Noam Brown on x.
Ray FernandoA developer and Early Adopter celebrated the drop in prices in a mail “LfG!” In short, let’s go! ”
The feeling reflects a growing enthusiasm of the builders who want to scale their projects without unaffordable access costs.
Price comparison with other competing LLMS
The price adjustment takes place at a time when AI providers compete more aggressive with performance and affordability. A comparison with other leading AI argumentation models shows how important this step could be:
- Gemini 2.5 per previewDeveloped by Google Deepmind, depending on the command prompt, loads between 1.25 and $ 2.50 for inputs and 10 to 15 US dollars for output. While integration in Google Search offers additional functions, this service offers its own costs – free of charge for the first 1,500 inquiries per day and then 35 USD per thousand.
- Closing work 4Marketed by Anthropic as a model that is optimized for complex tasks is the most expensive in the group and calculates $ 15 per million input token and 75 US dollars for production. Immediate caching reading and writing services is $ 1.50 or $ 18.75, although users can unlock a discount of 50% when processing.
- The models from Deepseek, especially Deepseek-R season and deepseek chat, undermine a large part of the market with aggressive low pricing. The at the entrance token range from $ 0.07 and 0.55, depending on the caching and time of day, while the output is between USD 1.10 and USD 2.19. The reduced prices during the off-peak lessons bring the prices even further to $ 0.035 for intermediate inputs.
| Model | Entrance | Intermediate entrance | output | Discount notes |
|---|---|---|---|---|
| Openai O3 | $ 2.00 (after $ 10.00) | $ 0.50 | $ 8,00 (after $ 40.00) | Flex processing: $ 5 / $ 20 |
| Gemini 2.5 Pro | $ 1.25 – $ 2.50 | $ 0.31 – $ 0.625 | $ 10.00 – $ 3,00 | A higher rate applies to input requests> 200k token |
| Closing work 4 | $ 15.00 | $ 1.50 (read) / $ 18.75 (writing) | $ 75.00 | 50% discount with batch processing |
| Deepseek-Chat | $ 0.07 (hit) $ 0.27 (Miss) | – | $ 1.10 | 50% discount while outside the peak times |
| Deepseek-Ried | $ 0.14 (hit) $ 0.55 (Miss) | – | $ 2.19 | 75% discount while outside the peak times |
In addition, the artificial analysis of the artificial analysis of AI model comparison and research group of third-party providers carried out the new O3 through its suite of benchmarking tests to various tasks and found that it cost $ 390 to complete them all, compared to $ 971 for Gemini 2.5 per and $ 342 for Claude 4.
Narrowing of the costs compared to intelligence gaps for developers
The pricing of Openaai not only narrows the gap with ultra-deep cost models such as Deepseek, but also has the pressure on higher-priced offers such as Claude Opus and Gemini Pro.
In contrast to Claude or Gemini, Openais O3 now also offers a flex mode for synchronous processing, in which 5 US dollars for input and 20 US dollars for output -Pro million tokens are calculated, which enables the developers more control over compensation costs and latency.
O3 is currently available via the Openai -API and the playground. Users with a weight of up to a few dollars can now examine the full skills of the model, which enables prototyping and deployment with fewer financial obstacles.
This could particularly benefit start-ups, research teams and individual developers who previously found more model access to cost-prohibitive.
Openai significantly lowers the costs for the most advanced argumentation model and signals a broader trend in the generative AI space: The premium performance quickly becomes more affordable, and the developers now have a growing number of sustainable, economically scalable options.
