Dev verdict on Google's new model: easier to work with than OpenAI or Anthropic, but odd mistakes

tokenbender · x · 2026-11-17

Developer tokenbender, after paying for an Ultra subscription and assessing the new model across his workflows, calls it a great release — easier to work with than the OpenAI or Anthropic counterparts — but notes it makes strange mistakes and, for some reason, irritates both OpenAI and Anthropic models when present in multi-model agent teams.

Original post →

More from Models

Models channel →