GPT4-0314 For the locally run model, we refer to the Language Model alone, not augmented with search/RAG/function_call. It needs a minimum throughput of 4 tokens/second Not sure what benchmarks people use in 2026, but let’s say LMSYS Arena for the moment. Will change depends on the trend. Current SOTA: I am not sure Phi3(3.8B) can fit on a phone. If not, the current bests are MiniCPM and Gemma 2B Update 2026-03-15 (PST) (AI summary of creator comment): The creator is resolving YES based on Gemma 3n being a phone-runnable model with an LMArena rating of 1308, which beats GPT-4-0314's rating of 1207. Update 2026-03-15 (PST) (AI summary of creator comment): The creator is resolving this market YES, citing Gemma 3n as a phone-runnable model with an LM Arena rating of 1308, which beats GPT-4-0314's rating of 1207.

Yes — resolved on Mar 16, 2026 by Manifold Markets prediction market.

MANIFOLD

Will a language model that runs locally on a consumer cellphone beat GPT4 by EOY 2026?

Ṁ1kṀ5.6k

resolved Mar 16

Resolved

YES

ALL

GPT4-0314

For the locally run model, we refer to the Language Model alone, not augmented with search/RAG/function_call. It needs a minimum throughput of 4 tokens/second

Not sure what benchmarks people use in 2026, but let’s say LMSYS Arena for the moment. Will change depends on the trend.

Current SOTA:

I am not sure Phi3(3.8B) can fit on a phone. If not, the current bests are MiniCPM and Gemma 2B

Update 2026-03-15 (PST) (AI summary of creator comment): The creator is resolving YES based on Gemma 3n being a phone-runnable model with an LMArena rating of 1308, which beats GPT-4-0314's rating of 1207.

Update 2026-03-15 (PST) (AI summary of creator comment): The creator is resolving this market YES, citing Gemma 3n as a phone-runnable model with an LM Arena rating of 1308, which beats GPT-4-0314's rating of 1207.

Market context

ChatGPT

GPT-4

Get

1,000

to start trading!

🏅 Top traders

#	Trader	Total profit
1		Ṁ460
2		Ṁ126
3		Ṁ81
4		Ṁ63
5		Ṁ57

People are also trading

Will we fully interpret a GPT-2 level language model by 2028?

14% chance

GPT-5 level model runnable on phones by 2030?

41% chance

Will $10,000 worth of AI hardware be able to train a GPT-3 equivalent model in under 1 hour, by EOY 2027?

17% chance

By 2028 will a language model beat the Ender Dragon?

54% chance

Before 2028, will anyone train a GPT-4-level model in a minute?

29% chance

Will a large language model beat a super grandmaster playing chess by EOY 2028?

60% chance

Will a large language model beat a super grandmaster at chess by EOY 2035?

Sort by:

Gemma 3n is a model runnable on phone with a higher LMArena rating of 1308, beating GPT-4-0314's 1207 rating. I am resolving yes.

@Sss19971997

@Sss19971997 hmmm interesting... my intuition would be that there's rating inflation in all Elo systems as more models are released! You can't just compare an Elo rating now with one from 2 years ago... but LMArena implies that the old rating is still active? Even though it's likely that no one had reviewed GPT-4 in years.

@bens The margin is big enough.

@bens No. why would there be inflation? it's ELo, it's relative, no such thing as inflation

Gemma 2 it 9b is already higher on llmsys arena. Only 2 points though.

Upd: oops sorry was thinking that criteria is <10b, confused with different question

bought Ṁ20 NO

I'm guessing this means any consumer cellphone? E.g. if a model that fits in 32GB RAM beats GPT-4 and there's only 1-2 phones with that much RAM in 2026 (current record is 24GB), this resolves Yes.

@JoshYou yes. Any consumer cellphone

bought Ṁ10 NO

Runs at what rate? Of its token per minute does it count?

@0482 Let’s say 4 tokens/s

if they are allowed to browse the internet - then for sure. If we are talking about encoding all the knowledge. Then probably not.

@Magnus_ Great question...

How should we specify this? I am thinking that it can do RAG for everything inside the phone but no internet connection. What do you think?

Currently, a phone can hold up to 512GB. It is a lot of info, but not the whole internet.

This criterion captures the "local" aspect.

how you think?

@Magnus_ Another option is to say language model only, no local RAG

@Magnus_ I thought about it again. For a fair comparison, we should have the same standard for the local mobile LM since GPT-4 is not using any RAG/search/tools. I have updated the criterion.