Content
@
0 reply
20 recasts
20 reactions
Buldevskaya
@buldevskaya
GPT-4o just got a serious upgrade. The model made big gains across all benchmarks — especially in coding. It’s now on par with the latest Claude 3.7 and updated DeepSeek V3. Still far behind the top-tier reasoners, but those are expensive and overkill for many tasks anyway. No official comparison with GPT-4.5 (benchmarking is $$$), but according to LMSYS Arena, the new GPT-4o is outperforming it.
0 reply
0 recast
1 reaction
