Content
@
https://opensea.io/collection/dev-21
0 reply
0 recast
2 reactions
Harpreet Singh
@happysingh
OS LLMs are just bad. I was using deepseek-v2-lite from @venice-ai cause it supports responseSchema and is meant for code For AI CLI Tool that sets up wallet integration in your Next Project web3wizard.dev For the same prompt, Deepseek was hallucinating and initially it took me a lot of time to configure the right prompt On the other hand, o4-mini was working perfect all the time. So, we need to have better version of those models.
0 reply
0 recast
0 reaction
