Opus 5.5 replaces Opus 5 at $4 / $20 per 1M tokens (vs $5 / $25), same context and output limits.
Should you care? Yes if you run Opus 5: ~20% cheaper. Budget a day for the four breaking changes before switching.
At a glance · vs Claude Opus 5
A developer ran a 744B-parameter AI (GLM 5.2) on a laptop without a GPU.
True, with a catchSpeed. On a 25 GB-RAM machine with a cold cache: ~0.05–0.1 tokens/sec, i.e. one token every 10–20 seconds. A 300-word answer can take over an hour.
The facts & how to try it →Tell us your models, cloud, region and frameworks. We flag only what affects you: new versions, price cuts, deprecations, breaking releases.
So what? A cheap open-weight option for high-volume classification on Bedrock. No structured-output support yet, so extraction pipelines need a validator.
So what? A runnable reranker that drops into existing retrieval pipelines. Worth a 30-minute test if your RAG answers cite the wrong page; our RAG suite run is scheduled.
So what? Nothing to do yet. We track it so you are not surprised; no facts will be added until a primary source confirms.