This market will resolve to "Yes" if OpenAI publicly releases the weights of a reasoning model (e.g., o1, o3) or a large language model (LLM) that is as advanced as GPT-3 by August 31, 2025, 11:59 PM ET. Otherwise, this market will resolve to “No”. Multimodal models (e.g., DALL·E, vision-language models), speech or audio-based models (e.g., Whisper), specialized AI agents or reinforcement learning models (e.g., robotics, AlphaCode-like systems) or models that otherwise do not meet the listed criteria will not qualify. The resolution source for this market will be a consensus of credible reporting.
[link removed] weights released, resolve?
GG
nobody resolve this market?
"gpt-oss is out! we made an open model that performs at the level of o4-mini and runs on a high-end laptop (WTF!!)" - @sama GG Boys
gg
You betting on the gpt-5 release day market?
no but if I did i would take thursday
[link removed]
Pump fake. Thanks for the discounts paper hands.
you really gonna hold this shit till it’s worthless?
it literally released, its called harmony, say goodbye to your position XD
[link removed] it's released. dump while u can.
oh well. thought it'd be gpt5 first. GG boyzzzz
Models have souls
Everything pointing to OSS model released today. Very clean market
Also funny how a lot of people have been citing news articles that always miss the mark, last friday it was supposed to be released "within hours" but we're still waiting lol, this market is free money: [link removed]
[link removed]
The release of open source AI model is imminent: [link removed]
"we have a ton of stuff to launch over the next couple of months--new models, products, features, and more. please bear with us through some probable hiccups and capacity crunches. although it may be slightly choppy, we think you'll really love what we've created for you!" - @sama on X. Nothing open source coming this month
Considering the imminent release of gpt-5, I doubt openai would want to fragment their focus and release an OS model during the same period. Openai also has no 'pressure' to release an OS model this soon. They haven't released an OS model since gpt-2, and that clearly hasn't hindered their growth in any way. Also, they wouldn't release a mediocre OS model just to get mogged by the Chienese qwen, kimi k2, or deepseek. I think they'll want to release a near-sota OS model, but that also means indefinite delays because of 'safety reasons' and 'further testing', as has already happened many times. 'Yes' shares are not crazy but I think they're overpriced. The only thing they have going for it is that more than a week ago, some gpt-codenamed configs for OS models were leaked on hugging face and quickly scrubbed, but other than that, there's been silence. Also, please do yourself a favor and do not listen to Dyogenes. If you scroll down the comments, you'll find him saying that he "doesn't know shit about AI" and also "didn't know what an LLM was" until a few days ago. Now he's trying to teach us lessons about GPUs and model performance in the comments lmao. What a joke
i totally understand being on the other side of this bet, but if youre genuinly buying this at 75 cents or higher, ur slow in the head
Good advice
80% is very underrated still, the guy promised before congress it would come in summer.
talking and doing are 2 different things, also, they can just release the 20B one for this summer as well as GPT5 and keep the 120 MoE in developement for a bit more.
which wouldn't be enought to finish the market in "yes" as for the 20B one so...
why wouldn't it ? reasoning model OR simple llm will do the trick for yes
or a large language model (LLM) that is as advanced as GPT-3 by August 31, 2025, 11:59 PM ET. Otherwise, this market will resolve to “No”.
GPT 3 is NOT a simple LLM
No you understand it wrong. OpenAI is gonna release a much better model than GPT-3 .. its obvious. GPT-3 is so old. And im fairly certain its gonna be next week.
it has nothing to do with being "old".... they announced a "O-3mini" like model, which is gonna be about 20B, and was made to be fast, and efficient to reduce cost per token, absolutely not to outperform GPT 3, i think that you didn't do any reshearch tbh
they also always had a policy of being aware of the AI, even an AI like GPT3 but optimised, would be way too powerful for the grand public, and all that in 20B ? very hard to believe, for them to release a "much better" model in open source that will reduce the use of their models, what you say make no sense, to you they would release an upgraded version of GPT 3 that was 175B, and an insane cost per token because of its power, reduced to 20B, more performant, faster, and accecible for all ? anf for FREE ? uhm... interresting point of view i must say
and they are releasing GPT 5 on top of that, so....
bro you are so behind on ai news. gpt-3 is garbage. there are 100s of chinese models surpassing it. o3-mini level is already 10x better xD, anyways your loss
whatever float you boats i guess.
He’s right o3 mini is more powerful than gpt3. Horizon alpha and beta are gpt5 variants which will definitely not be open source; they ain’t giving that away for free. They’re under negotiations with Microsoft, they want to covert to a for profit company, which means they need to raise/generate capital.
damn, i have been reported
the truth ? i don't even know shit about Ai's, didn't knew tf was an LLM till yesterday
saw a scalping opportunity and took it that's all. But i guess, as long as you mix some truths and you say it with enought confidence it become reality for some, or perhaps i sisn't said only bs, either way i supported the movement. I took my profits, it's time for me to move out, UNTIL WE MEET AGAIN !
I’m in this for the long run til 2025 🤪. Nothing wrong with scalping. ✌🏻
hey ! it's me again. time for a little lesson, open sources model are primarily made to be able to run locally. a 4090 which is one of the best grapgic card can barely run a 30B AI, now i don't think a lot of you read the actual rules of the market. So it's either they make one of their reasonning language public, or they make a new one as powerful as GPT3, which is 175B, so impossible to run locally therefore it would be pointless. So, basicly, even if they do invent a
new one, it will need to be aprox 150B minimum, so impossible to run locally, or they can an existing one public, which is higly unlikely, it would be seen as recycling
buy me all at 75.6
nah, i won't rob you of 25% bro, what a bad guy would i be right ? your so close to winning 1.3K !
For everyone listening to poker brat... just so you know, they are planning to release a model similar to o-3 mini which is way less powerful than GPT 3, even IF a model is released in August, it won't come as close as davinci GPT-3 you are welcome
Davinci is 3 years old bruh
and ?
you really are restarted
GPT 3 was 175B, o-3 mini is between roughly 10 to 20B MAX. It was made to reduce the cost per token, at the cost of his power, and since they are planing on releasing a O3-mini like, then it won't go above 20B or so, which is way faster, but also way less powerful
and good luck running a 175 B locally, which is the principal arguments for open sources.
That guy is the most unreliable puss handed faker here.
Next week boyos
Sold and was waiting for a price drop but they leaked the fucking weights
link ?
they leaked repo names.... Also ! It's talking about o-3 mini..... not even close from GPT-3 power. sold low, bought high. a real example
[link removed]
Also one guy got the weights before they noticed lol
still, even the biggest model is 120B, i doubt they release the 2 together, that is an awful lot to be an open one, maybe made for specific function ? but open source at 120B is not for private useage, and even tho it was, still need the benchmark, that match the 175B of GPT3
that is 55B on missing information, and most likely due to voluntarily nerf the system, they had always a take on AI and the fact that it must be given with great responsability. so the missing infos might be advanced reasoning
03 mini is more powerful than gpt3.. gpt5 is what’s coming out. Horizon alpha and beta are gpt5 variants.
Open source ain’t coming out. They’re more focused on generating money, that’s why they are having intense negotiations with Microsoft so that they can convert to a for profit
On Capacity Issues (April 2025): "We are getting things under control, but you should expect new releases from OpenAI to be delayed, stuff to break, and for service to sometimes be slow as we deal with capacity challenges" (Startup News) (Sam Altman) and "Working as fast [as] we can to really get stuff humming." (Startup News).-Scam Altman
On the Open-Source Model Delays (July 2025): (Yahoo Finance) (TechCrunch) "We need time to run additional safety tests and review high-risk areas. we are not yet sure how long it will take us" and "while we trust the community will build great things with this model, once weights are out, they can't be pulled back. this is new for us and we want to get it right."-Scam Altman
120b model
120b moe 20b sparse