- The Pulse by 42neurons
- Posts
- GPT-4.5 preview's accuracy on GPQA Diamond
GPT-4.5 preview's accuracy on GPQA Diamond
Hey there!
Welcome back to The Pulse, where we dive into interesting AI stories and trends backed by data, all presented through simple visuals.

> GPT-4.5 preview scores highest accuracy among non-reasoning models
> proves that scaling law still holds: bigger = smarter models

> Claude 3.7 Sonnet shows best returns
> OpenAI models more expensive than the rest
> GPT 4.5 preview ~x500 more expensive than DeepSeek Chat V3

> n= 750 C-suite executives from American & European companies with revenues >$100M
> 88% see positive impact on revenue growth; 37% consider that impact significant
> companies with revenues under $500M are less optimistic than bigger ones
> American, British and French companies more skeptical than others
.