Will an LLM with less than 10B parameters beat GPT4 by EOY 2025?

260Ṁ1812

resolved Jan 10

Resolved

YES

ALL

How much juice is left in 10B parameters?

the original GPT4-0314 (ELO 1188)

judged by Lmsys Arena Leaderboard

current SoTA: Llama 3 8B instruct (1147)

Technical AI Timelines

LLMs

GPT-4 speculation

Get

1,000

to start trading!

🏅 Top traders

#	Name	Total profit
1		Ṁ47
2		Ṁ44
3		Ṁ32
4		Ṁ26
5		Ṁ23

People are also trading

Will the best LLM in 2025 have <1 trillion parameters?

42% chance

Will the best LLM in 2025 have <500 billion parameters?

23% chance

Will the best LLM in 2026 have <1 trillion parameters?

40% chance

Will the best LLM in 2027 have <1 trillion parameters?

26% chance

Size of smallest open-source LLM marching GPT 3.5's performance in 2025? (GB)

1.83

Will the best LLM in 2026 have <500 billion parameters?

27% chance

China will make a LLM approximately as good or better than GPT4 before 2025

89% chance

Will the best LLM in 2027 have <500 billion parameters?

13% chance

Will the best LLM in 2027 have <250 billion parameters?

Sort by:

bought Ṁ500 YES

@mods Resolves as YES.. The creator's account is deleted, but Gemini 1.5 Flash-8B easily clears the bar (ELO higher than 1188) with an ELO of 1212, see https://lmarena.ai/

bought Ṁ400 YES

The market creator is banned, but it seems the YES resolution criterion has been met because Gemma-2-9B-it is at elo 1188, above GPT-4-0314 at 1886. However, one could argue it needs to be 1189 or higher because GPT-4-0314 was at 1188 when the market was created.

bought Ṁ10 YES

Open source models have an inherent advantage in that they don't need to conform to censorship/"safety" policies, making them much more useful and thus getting them a higher ELO. Having to conform to policies is a major handicap for the larger closed-source models.