I think the clue is hard to miss.
Will there be more?
Event starts at:
2025-04-16T17:00:00Z
This topic will be updated as the event unfolds!
…
Here is the link to the YouTube stream:
Thanks @polepole !
Comments during the event will be deactivated. Share your impressions and thoughts here instead!
Looking forward to it!
They are unbagging cats in the title of the video…
GPT-4.1 was quite resistant to turning a script into a non-crashing UI that also functions. Probably trained too much on style sheets…
GDB demoing again after 2 years.
2025-04-16 17:25:00 UTC: o3
2025-04-16 17:25:00 UTC: o3-2025-04-16
2025-04-16 17:25:00 UTC: o4-mini
2025-04-16 17:25:00 UTC: o4-mini-2025-04-16
Codex CLI client from OpenAI!!!
System card for the new models:
Page on thinking with images:
https://openai.com/index/thinking-with-images/
API pricing for new models:
| Model | Description | Input/Mtok | Cached Input/Mtok | Output/Mtok |
|---|---|---|---|---|
| o3 | Our most powerful reasoning model with leading performance on coding, math, science, and vision | $10.00 | $2.50 | $40.00 |
| o4-mini | Our faster, cost-efficient reasoning model delivering strong performance on math, coding and vision | $1.100 | $0.275 | $4.40 |
https://cdn.openai.com/pdf/2221c875-02dc-4789-800b-e7758f3722c1/o3-and-o4-mini-system-card.pdf
BONUS:
Finally, the much-awaited moment of ultimate naming conflict has arrived: 4o vs. o4.
4o-mini vs. o4-mini.
I still can’t understand the terrible idea of naming the reasoning models with an “o” instead of an “r,” especially considering they’ve already used “o” in the GPT family. 
The new “verify your ID” now required on these models for:
-Any access to tier 1-3
While the verification help page claims the need is because of input abuse…
It appears the documentation is still being updated regularly. I expect the most up to date information will always be available here:
https://help.openai.com/en/articles/10910291-api-organization-verification
Besides the fact that it was really cool to see o3 in action, there are two other fantastic details worth pointing out:
The costs for o3 are much—repeat, much—lower than anticipated.
There’s no indefinite waiting period for the EU and other countries. It has already arrived in the API and ChatGPT
Tthis is rather fundamental information to present when finding this forum topic - and thinking you’d be able to employ the API reference:
HTTP error occurred: 400 - {
"error": {
"message": "Your organization must be verified to generate reasoning summaries. Please go to: https://platform.openai.com/settings/organization/general and click on Verify Organization.",
"type": "invalid_request_error",
"param": "reasoning.summary",
"code": "unsupported_value"
}
}
Error: Client error '400 Bad Request' for url 'https://api.openai.com/v1/responses'
For more information check: https://developer.mozilla.org/en-US/docs/Web/HTTP/Status/400
O3 is simply - no streaming without verification. Becuase of the trust required in doing so?
Edit: please refer to the official help pages for all questions regarding this matter.
The reply below it’s outdated and by now incorrect.
https://help.openai.com/en/articles/10910291-api-organization-verification
Mea culpa, I need to specify this information:
Tier 4 and 5 will not have to verify but Tier 1-3 must do so.
So it’s an Agent wrapper?
Verified API access for Tier 1 , specifically o4 mini - o4-mini-2025-04-16. I am running evals and still early thoughts/data points but this could really impact RAG use cases and alleviate citation hallucinations. Organization Verification is not enforced, perhaps should be enforced.
o4-mini-high huge downgrade
Is there any plans to enable fine-tuning of reasoning model?