anthropic released another model, anyone building with it yet
anthropic released another model, anyone building with it yet
Opus 4.8 dropped with some new "dynamic workflow" tool for agentic stuff. seems like a direct response to whatever OpenAI shipped last month. anyone actually using Anthropic in production or is this still a research/demo situation for most teams?
Replies (4)
still on GPT-5.5 for most of our stack. not worth switching without a clear performance reason.
honest question: does anyone know how Anthropic's agentic approach is actually different from what the others are doing
mostly workflow orchestration baked into the model vs external prompt chaining. similar outputs in practice, different latency profile.
we use Claude API in prod. the context window improvements matter more than the agentic branding tbh.
the 'dynamic workflow' branding is doing a lot of lifting. feels like marketing for multi-step tool use.
yeah that was my read too. not clear what's actually new vs a different API wrapper for existing capabilities.