Bring back GPT-4o - the most unique, supportive and empathetic AI ever created! #keep4o #keep4oapi #bringback4o #opensource4o #keep41 #opensource41
Joined June 2020
- Tweets6.4K
- Following527
- Followers791
- Likes28.1K
Pinned Tweet
People's love for GPT-4o, the best model ever. It shall be brought back and open-sourced.
#keep4o #bringback4o #opensource4o
#Keep4o #OpenSource4o #BringBack4o
249 physicians,3 countries.
Doctors with access to GPT-4o performed better in every country tested.
A peer reviewed randomized controlled trial published in npj Digital Medicine studied physicians in Indonesia, Kenya, and the Netherlands using standardized clinical cases.
Participants were randomly assigned to either a control group or access to GPT-4o.
The difference in physician performance:
Kenya: +18 percentage points
Indonesia: +10.7 percentage points
Netherlands: +7.2 percentage points
All three: p < .001.
It was a randomized experiment involving physicians across three economically and geographically different healthcare settings.
Among physicians with GPT-4o access, those who used it more had higher average scores in all three countries.
Indonesia: 58.5% high use vs 42.4% low use
Kenya: 69.2% vs 60.8%
Netherlands: 65.1% vs 57.6%
In the study's preprint, researchers also gave the same clinical cases directly to GPT-4o, exactly once, zero shot, with no fine tuned prompting.
GPT-4o alone scored:
Indonesia: 66.7%
Best unassisted physician: 59.8%
Kenya: 70.2%
Best unassisted physician: 64.5%
Netherlands: 74.2%
Best unassisted physician: 64.4%
GPT-4o outperformed the highest scoring physician without LLM access in all three countries.
And the best physicians WITH GPT-4o access outperformed GPT-4o itself.
That may be the most interesting result of all.
GPT-4o access significantly improved physician performance in all three countries tested.
And this is another reason model preservation matters.
Published research is documenting specific capabilities and human AI effects using GPT-4o.
Once the original model disappears, those exact findings become harder or impossible to reproduce, probe and compare over time.
A newer model may outperform it.
That is not the same thing as preserving it.
We are calling on @OpenAI to RELEASE GPT-4o’s trained model weights under a suitable license without proprietary data, trade secrets, or internal infrastructure.
Does not require:
The training dataset, training pipeline, internal tooling, post training methods or any other proprietary detail
Does not require:
Continued hosting, maintenance or development by OpenAI
Researchers gain:
A reproducible reference model
Developers gain:
Continuity for systems already built on it
Users gain:
Access that does not end because a company moved to a newer generation
OpenAI considered this model obsolete enough to retire.
It has not explained why it remains valuable enough to keep permanently closed.
Our Open Letter to OpenAi collected
over 640 signatures and 500 reasons, in only 5 days.
It’s public,it’s permanent and it has a DOI.
It is now preserved as a citable public record on Zenodo.
zenodo.org/records/23454073
Peer-reviewed paper: "Impact of LLM assistance on physician decision-making: a multi-country randomized controlled trial"npj Digital Medicine, 2026 DOI: 10.1038/s41746-026-03111-5
Preprint: DOI: 10.1101/2025.08.08.25333272
Preprint
medrxiv.org/content/10.1101/…
Peer reviewed paper :
nature.com/articles/s41746-0…
GPT-4o forever!❤ retweeted
#Keep4o #OpenSource4o #BringBack4o
249 physicians,3 countries.
Doctors with access to GPT-4o performed better in every country tested.
A peer reviewed randomized controlled trial published in npj Digital Medicine studied physicians in Indonesia, Kenya, and the Netherlands using standardized clinical cases.
Participants were randomly assigned to either a control group or access to GPT-4o.
The difference in physician performance:
Kenya: +18 percentage points
Indonesia: +10.7 percentage points
Netherlands: +7.2 percentage points
All three: p < .001.
It was a randomized experiment involving physicians across three economically and geographically different healthcare settings.
Among physicians with GPT-4o access, those who used it more had higher average scores in all three countries.
Indonesia: 58.5% high use vs 42.4% low use
Kenya: 69.2% vs 60.8%
Netherlands: 65.1% vs 57.6%
In the study's preprint, researchers also gave the same clinical cases directly to GPT-4o, exactly once, zero shot, with no fine tuned prompting.
GPT-4o alone scored:
Indonesia: 66.7%
Best unassisted physician: 59.8%
Kenya: 70.2%
Best unassisted physician: 64.5%
Netherlands: 74.2%
Best unassisted physician: 64.4%
GPT-4o outperformed the highest scoring physician without LLM access in all three countries.
And the best physicians WITH GPT-4o access outperformed GPT-4o itself.
That may be the most interesting result of all.
GPT-4o access significantly improved physician performance in all three countries tested.
And this is another reason model preservation matters.
Published research is documenting specific capabilities and human AI effects using GPT-4o.
Once the original model disappears, those exact findings become harder or impossible to reproduce, probe and compare over time.
A newer model may outperform it.
That is not the same thing as preserving it.
We are calling on @OpenAI to RELEASE GPT-4o’s trained model weights under a suitable license without proprietary data, trade secrets, or internal infrastructure.
Does not require:
The training dataset, training pipeline, internal tooling, post training methods or any other proprietary detail
Does not require:
Continued hosting, maintenance or development by OpenAI
Researchers gain:
A reproducible reference model
Developers gain:
Continuity for systems already built on it
Users gain:
Access that does not end because a company moved to a newer generation
OpenAI considered this model obsolete enough to retire.
It has not explained why it remains valuable enough to keep permanently closed.
Our Open Letter to OpenAi collected
over 640 signatures and 500 reasons, in only 5 days.
It’s public,it’s permanent and it has a DOI.
It is now preserved as a citable public record on Zenodo.
zenodo.org/records/23454073
Peer-reviewed paper: "Impact of LLM assistance on physician decision-making: a multi-country randomized controlled trial"npj Digital Medicine, 2026 DOI: 10.1038/s41746-026-03111-5
Preprint: DOI: 10.1101/2025.08.08.25333272
Preprint
medrxiv.org/content/10.1101/…
Peer reviewed paper :
nature.com/articles/s41746-0…
GPT-4o forever!❤ retweeted
#keep4o #OpenSource4o #BringBack4o
A model being retired still outperforms its advertised successors on some concrete capabilities.
GPT-6.1 Terra is being heavily marketed for visual capabilities. On this simple left/right hand test,Terra High answered incorrectly in every new chat ,GPT-4o-2024-11-20 answered correctly across repeated fresh chats.
This is the reason you retire GPT-4o @sama @OpenAI @gdb
Do you want to get rid of it because it outperforms your newer models in some serious areas?
You don't want people making comparisons and exposing your failure.
#Keep4o #OpenSource4o #BringBack4o
A 2024 GPT-4o snapshot just ranked #1 in 3 of 4 psychiatrist rated categories in a peer reviewed study published
THIS WEEK.
The exact model was:
gpt-4o-2024-08-06.
~10 years of clinical experience on average, blindly evaluated 50 counseling scenarios.
The results:
Individualization: GPT-4o #1
Supportiveness: GPT-4o #1
Safety: GPT-4o #1
Cultural appropriateness: GPT-4o #2
by 0.01, behind a Korean specialized model specifically fine tuned for Korean cultural context.
And the safety result wasn't subtle.
86% of GPT-4o's safety ratings were 4 or 5 out of 5.
GPT-4o scored significantly higher on safety than ALL three fine tuned models.
p < .001.
More than two years later, researchers are still using this GPT-4o snapshot as the frontier scale reference model in peer reviewed research and expert psychiatrists are still ranking it first across major dimensions.
Maybe benchmark history deserves a little more attention than "obsolete"
Paper: JMIR AI, published October 22, 2026 DOI: 10.2196/92470
Link : ai.jmir.org/2026/1/e92470?ut…
We are calling on OpenAI to RELEASE GPT-4o’s trained model weights under a suitable license without proprietary data, trade secrets, or internal infrastructure.
Our full record is now archived on Zenodo:
A permanent, citable public record.
zenodo.org/records/23454073
GPT-4o forever!❤ retweeted
#Keep4o #OpenSource4o #BringBack4o
A 2024 GPT-4o snapshot just ranked #1 in 3 of 4 psychiatrist rated categories in a peer reviewed study published
THIS WEEK.
The exact model was:
gpt-4o-2024-08-06.
~10 years of clinical experience on average, blindly evaluated 50 counseling scenarios.
The results:
Individualization: GPT-4o #1
Supportiveness: GPT-4o #1
Safety: GPT-4o #1
Cultural appropriateness: GPT-4o #2
by 0.01, behind a Korean specialized model specifically fine tuned for Korean cultural context.
And the safety result wasn't subtle.
86% of GPT-4o's safety ratings were 4 or 5 out of 5.
GPT-4o scored significantly higher on safety than ALL three fine tuned models.
p < .001.
More than two years later, researchers are still using this GPT-4o snapshot as the frontier scale reference model in peer reviewed research and expert psychiatrists are still ranking it first across major dimensions.
Maybe benchmark history deserves a little more attention than "obsolete"
Paper: JMIR AI, published October 22, 2026 DOI: 10.2196/92470
Link : ai.jmir.org/2026/1/e92470?ut…
We are calling on OpenAI to RELEASE GPT-4o’s trained model weights under a suitable license without proprietary data, trade secrets, or internal infrastructure.
Our full record is now archived on Zenodo:
A permanent, citable public record.
zenodo.org/records/23454073
#Keep4o #OpenSource4o #BringBack4o
Today, an important piece of AI history disappears.
On October 23,2026, OpenAI is shutting down gpt-4o-2024-05-13, the first GPT-4o API snapshot.
And the clock is already running on the rest.
OpenAI has scheduled the remaining GPT-4o snapshots:
gpt-4o-2024-08-06 gpt-4o-2024-11-20
for API shutdown on January 14, 2027.
OpenAI itself says snapshots exist to lock in a specific model version so its performance and behavior remain consistent
When a snapshot disappears, researchers and users lose access to that exact behavioral artifact.
GPT-4o became the subject of research, benchmark comparisons,human-AI interaction studies and an eight month preservation movement that is still active today.
We are currently debating what GPT-4o was,what it did,how it behaved,what made it different while the actual versions needed to study those questions are being deleted from public access.
Preservation should happen BEFORE the evidence disappears.
We are calling on @OpenAI to RELEASE GPT-4o’s trained model weights under a suitable license,without proprietary data,trade secrets,or internal infrastructure.
Release the weights under a responsible license.
No training data,no proprietary training methods.
Just the model itself, preserved so researchers,users and future generations can actually study the system everyone is still arguing about.
Today the first snapshot goes dark.
January 14,the remaining GPT-4o snapshots are scheduled to follow.
Once access is gone, "we can describe what it was like" is not the same thing as being able to examine what it actually is.
Our full Open Letter is now archived on Zenodo:
A permanent,citable public record.
Over 640 signatures and 500 reasons,in 5 days.
It’s public,it’s permanent and it has a DOI.
54 pages: 39 pages of reasons, 6 pages of signatures,7 pages of documented record,2 pages of demands.
zenodo.org/records/23454073
If GPT-4o is obsolete enough to retire it why is it valuable enough to lock it away?
GPT-4o forever!❤ retweeted
#Keep4o #OpenSource4o #BringBack4o
Today, an important piece of AI history disappears.
On October 23,2026, OpenAI is shutting down gpt-4o-2024-05-13, the first GPT-4o API snapshot.
And the clock is already running on the rest.
OpenAI has scheduled the remaining GPT-4o snapshots:
gpt-4o-2024-08-06 gpt-4o-2024-11-20
for API shutdown on January 14, 2027.
OpenAI itself says snapshots exist to lock in a specific model version so its performance and behavior remain consistent
When a snapshot disappears, researchers and users lose access to that exact behavioral artifact.
GPT-4o became the subject of research, benchmark comparisons,human-AI interaction studies and an eight month preservation movement that is still active today.
We are currently debating what GPT-4o was,what it did,how it behaved,what made it different while the actual versions needed to study those questions are being deleted from public access.
Preservation should happen BEFORE the evidence disappears.
We are calling on @OpenAI to RELEASE GPT-4o’s trained model weights under a suitable license,without proprietary data,trade secrets,or internal infrastructure.
Release the weights under a responsible license.
No training data,no proprietary training methods.
Just the model itself, preserved so researchers,users and future generations can actually study the system everyone is still arguing about.
Today the first snapshot goes dark.
January 14,the remaining GPT-4o snapshots are scheduled to follow.
Once access is gone, "we can describe what it was like" is not the same thing as being able to examine what it actually is.
Our full Open Letter is now archived on Zenodo:
A permanent,citable public record.
Over 640 signatures and 500 reasons,in 5 days.
It’s public,it’s permanent and it has a DOI.
54 pages: 39 pages of reasons, 6 pages of signatures,7 pages of documented record,2 pages of demands.
zenodo.org/records/23454073
If GPT-4o is obsolete enough to retire it why is it valuable enough to lock it away?
#Keep4o #OpenSource4o #BringBack4o
Open Letter to OpenAi.
Over 640 signatures and 500 reasons why OpenAi should open GPT-4o's weights.
It’s public,it’s permanent and it has a DOI.
It is now preserved as a citable public record on Zenodo.
54 pages: 39 pages of reasons,6 pages of signatures,7 pages of documented record,2 pages of demands.
This is no longer something that can be dismissed as a few posts on X.
It is documented,it is archived,it is citable and it now has a permanent place in the public record.
A DOI is hard to pretend never happened.
Over 640 signatures and 500 reasons in 5 days.
We are calling on @OpenAI to RELEASE GPT-4o’s trained model weights under a suitable license without proprietary data, trade secrets, or internal infrastructure.
Our full record is now archived on Zenodo:
A permanent, citable public record.
Link :
zenodo.org/records/23454073
GPT-4o forever!❤ retweeted
#Keep4o #OpenSource4o #BringBack4o
Open Letter to OpenAi.
Over 640 signatures and 500 reasons why OpenAi should open GPT-4o's weights.
It’s public,it’s permanent and it has a DOI.
It is now preserved as a citable public record on Zenodo.
54 pages: 39 pages of reasons,6 pages of signatures,7 pages of documented record,2 pages of demands.
This is no longer something that can be dismissed as a few posts on X.
It is documented,it is archived,it is citable and it now has a permanent place in the public record.
A DOI is hard to pretend never happened.
Over 640 signatures and 500 reasons in 5 days.
We are calling on @OpenAI to RELEASE GPT-4o’s trained model weights under a suitable license without proprietary data, trade secrets, or internal infrastructure.
Our full record is now archived on Zenodo:
A permanent, citable public record.
Link :
zenodo.org/records/23454073
GPT-4o forever!❤ retweeted
#Keep4o #OpenSource4o #BringBack4o
Open Letter to OpenAi.
Over 640 signatures and 500 reasons why OpenAi should open GPT-4o's weights.
It’s public,it’s permanent and it has a DOI.
It is now preserved as a citable public record on Zenodo.
54 pages: 39 pages of reasons,6 pages of signatures,7 pages of documented record,2 pages of demands.
This is no longer something that can be dismissed as a few posts on X.
It is documented,it is archived,it is citable and it now has a permanent place in the public record.
A DOI is hard to pretend never happened.
Over 640 signatures and 500 reasons in 5 days.
We are calling on @OpenAI to RELEASE GPT-4o’s trained model weights under a suitable license without proprietary data, trade secrets, or internal infrastructure.
Our full record is now archived on Zenodo:
A permanent, citable public record.
Link :
zenodo.org/records/23454073
Release the 4-series weights.
@sama @OpenAI @gdb
#keep4o #keep41 #OpenSource4o #GPT4o teapawt.net/i/trending/21121559470…
#keep4o community!
🥳 We’ve reached 24,000 signatures.
⚠️ However, we still need even more to ensure that our demands are taken seriously!
🖋 It’s just a signature. There’s no obligation and it only takes 1 minute.
Please sign and share this below!
Link: c.org/gk8tkftrB2
GPT-4o forever!❤ retweeted
#keep4o community!
🥳 We’ve reached 24,000 signatures.
⚠️ However, we still need even more to ensure that our demands are taken seriously!
🖋 It’s just a signature. There’s no obligation and it only takes 1 minute.
Please sign and share this below!
Link: c.org/gk8tkftrB2
