Sonnet 4.5 is now available for all users in Cursor as regular and thinking model.
Let us know how it works for you.
Sonnet 4.5 is now available for all users in Cursor as regular and thinking model.
Let us know how it works for you.
Even though it’s only recently released, the cursor team has added it very quickly. From what I can see, the processing time seems to be fast as well; less thinking, more examining.
Claude Sonnet 4.5 \ Anthropic : Sonnet 4.5 is the best model in the world for agents, coding, and computer use.
Yeah, but benchmarks show that gpt-5-high has a higher score. I am in doubt, going to test this model.
Sonnet 4.5 is now available for all users in Cursor as regular and thinking model.
Let us know how it works for you.
It is interesting that they included it in the free plan and the fee for API usage is the same as Sonnet 4.
@matgenehr which benchmarks show that?
OpenAI reports GPT‑5 at 74.9% on SWE‑bench Verified and 88% on Aider Polyglot, indicating strong end‑to‑end coding execution and tool‑use improvements for multi‑step tasks.
Anthropic reports Claude Sonnet 4 at ≈72.7% on SWE‑bench, placing it within a few points of GPT‑5; independent roundups echo a slim gap in coding, with Sonnet favored for efficiency and stability in some workflows.
Recent community and vendor leaderboards list Claude 4 Sonnet and Opus in the low‑70s on SWE‑bench, consistent with Anthropic’s announcement and third‑party aggregations.
Those are benchmarks for Claude 4.
From what I’ve seen, Claude 4.5 does a lot better
I hope they provide free usage.
Cursor administrators can conduct a one-week trial with generous donations.
Have you tested both?
I am willing to test, but I don’t want it to screw my database or make changes to my code that I have to work overhours fixing.
And gpt is working well for me.
Does Claude 4.5 spend money at the same rate as 4.1?
2x requests, eh?
Will Claude Sonnet 4.5 be available in Auto Mode? I noticed that the cost of Claude Sonnet 4.5 is the same as Claude Sonnet 4, and it seems to serve as its full successor. However, Claude Sonnet 4.5 is described as a “thinking” model. Unfortunately, Auto Mode very rarely includes models with deeper reasoning capabilities.
right now, really fast. beats codex in precision and speed.
@condor Isn’t Cursor affiliated with Anthropic? Can’t Cursor lobby them for a free week (or two!!) for Cursor users?