What will be true of OpenAI’s open-weight model?

Ṁ4993

Dec 31

ALL

90%

At least 20B Parameters

29%

released in September, October, November, or December 2025

68%

released in the month of August 2025

released in the month of July 2025

released in the month of June 2025

96%

Is a reasoning model

Is at least an Image-Language Model (ala JanusPro or 4o)

86%

Is at least a Vision-Language Model (ala PaliGemma)

80%

Uses Beyer Teacher/Student distillation ala Gemma 3

11%

my friend @soaffine (who knows ab ai) thinks it unveils an impressive new architecture

73%

An instruct model is made accessible

18%

A base model is made accessible

45%

The model is accessible to all users on the main chatgpt site

20%

Gets 1400+ Elo on lmarena.ai

14%

At least 300B parameters

53%

At least 70B parameters

68%

At least 30B parameters

67%

better than o3-mini on FrontierMath (with tools)

84%

Uses Mixture of Experts

60%

MIT or Apache license

Context:

Feel free to add your own answers. If an answer is unclear you are welcome to ask for clarification from both me and the person who submitted the answer.

#AI

#Technical AI Timelines

#OpenAI

#AI Impacts

#LLMs

Get

1,000

and

1.00

6 Comments

Sort by:

"reasoning model" will be judged with reference to OpenAI documentation.

bought Ṁ10 Is at least an Image... YES

What’s the difference between image/language and vision/language?

bought Ṁ10 Is at least an Image... NO

@KJW_01294 I think image/language generates images whereas vision/language only takes them as input

I've chosen to bet on the market that references me, but assert that I will be truthful in my evaluation of it. To help others, I would have said no to Gemma 2/3, yes to PaliGemma, yes to DeepSeek-MoE, but no to DeepSeek V3 and DeepSeek R1

bought Ṁ10 MIT or Apache license YES

"parameters" means total parameters?

bought Ṁ30 Uses Mixture of Experts NO

@JoshYou correct

Related questions

Related questions