How much FLOP will be used to train the best language model with freely available weights on July 1, 2025? | Manifold

How much FLOP will be used to train the best language model with freely available weights on July 1, 2025?

5

535Ṁ1237

Jul 2

2%

<1e24

2%

[1e24, 3e24)

12%

[3e24, 1e25)

15%

[1e25, 3e25)

33%

[3e25, 1e26)

25%

[3e26, 1e27)

4%

[1e27, 3e27)

1.9%

[3e27, 1e28)

1.6%

[1e28, 3e28)

1.2%

[3e28, 1e29)

1.2%

>=1e29

Get

1,000

to start trading!

Sort by:

There seems to be a mistake, there is no option if someone believes it will be 2e26.

How would the best language model be determined?

People are also trading

Will an AI model use more than 1e28 FLOPS in training before 2026?

End of pre-training era for language models: Will an LM fine-tune for more FLOPs than it is pre-trained for, before 2026

Will a machine learning training run exceed 10^26 FLOP in China before 2026?

Will the largest machine learning training run (in FLOP) as of the end of 2025 be in the United States?

Will a lab train a >=1e26 FLOP state space model before the end of 2025?

Will there be an announcement of a model with a training compute of over 1e30 FLOPs by the end of 2025?

Will a flagship (>60T training bytes) open-weights LLM from Meta which doesn't use a tokenizer be released in 2025?

Will a machine learning training run exceed 10^25 FLOP in China before 2027?

Most training run compute greater than 2e27 FLOP by EOY 2026?

Will any 10 trillion+ parameter language model that follows instructions be released to the public before 2026?

Related questions

Will an AI model use more than 1e28 FLOPS in training before 2026?

End of pre-training era for language models: Will an LM fine-tune for more FLOPs than it is pre-trained for, before 2026

Will a machine learning training run exceed 10^26 FLOP in China before 2026?

Will the largest machine learning training run (in FLOP) as of the end of 2025 be in the United States?

Will a lab train a >=1e26 FLOP state space model before the end of 2025?

Will there be an announcement of a model with a training compute of over 1e30 FLOPs by the end of 2025?

Will a flagship (>60T training bytes) open-weights LLM from Meta which doesn't use a tokenizer be released in 2025?

Will a machine learning training run exceed 10^25 FLOP in China before 2027?

Most training run compute greater than 2e27 FLOP by EOY 2026?

Will any 10 trillion+ parameter language model that follows instructions be released to the public before 2026?

© Manifold Markets, Inc.•Terms•Privacy