The author comments on the release of Anthropic's Opus 5.5 model, expressing reduced excitement and a growing preference for open-source AI models from companies like Alibaba and DeepSeek, highlighting a shift in interest within the AI community.
Not even that long ago I was constantly waiting for the next Anthropic model. New Claude announcement used to be one of those “ok drop everything, gotta see the benchmarks” moments for me. Opus 5.5 came out like a couple hours ago and my reaction was basically …cool? And it’s not even that the model looks bad. It doesn’t. You are getting Fable like performance for fraction of the cost and all. I just realized I’m way more excited these days when I see Qwen, MiMo, DeepSeek, GLM etc releasing something. A year or two ago I wouldn’t have battered a second eye to these companies. Now Alibaba casually drops jawdropping stuff that you can actually download, quantize, run on your own hardware, mess with MTP, modify inference, fine tune, whatever… and that’s somehow way more interesting to me than another closed frontier model getting better at coding. Obviously Opus is still going to destroy local models on plenty of tasks. Not really claiming otherwise. It’s more that the interesting part of AI for me has shifted.l which is honestly kinda crazy considering where things were a couple years ago? Anyway just wanted to rant a bit I guess. Do you guys share the smae feeling?
Anthropic has released Opus 5.5, a new AI model with lower prices and performance matching or exceeding larger models like Fable, featuring improved communication and alignment with safety pacing efforts.