A new Grok model, grok-code-fast-1, is now available in Cursor! It is fast and has competitive pricing ($0.2/1M input, $1.5/1M output). It is free for one week. Feel free to try it out! Source

We've seen bunch of AI companies release a CLI coding tool and now Cursor also joined the party It supports all the same models as we've used to in Cursor The CLI also supports MCP, Rules, and everything else we already have in the IDE The CLI is in Beta right now You can install it by running curl https://cursor.com/install -fsS | bash Then try it out by running cursor-agent Note: This is directly from Cursor docs but I have NOT verified the script. Run at your own risk! Check the docs & how to use the CLI here: https://docs.cursor.com/en/cli/overview

TL;DR: - GPT-5 is great for coding - For free in Cursor for a week Full story: OpenAI just released their latest and greatest flagship model GPT-5 The tagline for the model is "the smartest, fastest, most useful model yet, with built-in thinking that puts expert-level intelligence in everyone’s hands" It has state-of-the-art performance across coding, math, writing, health, visual perception, and more Instead of being a model like we've used to, GPT‑5 is a unified system with a smart, efficient model that answers most questions, a deeper reasoning model (GPT‑5 thinking) for harder problems, and a real‑time router that quickly decides which to use based on conversation type, complexity, tool needs, and more. GPT-5 is also the strongest coding model from OpenAI to date. In particular, it exceeds in front-end generation and it can create beautiful and responsive websites, apps, and games with way better eye for aesthetics and design than previous models. In coding benchmarks, it beats o3, the previous best model from OpenAI (SWE-bench and Aider Polyglot) In LiveCodeBench and SciCode, it beats Claude 4 Sonnet and slightly loses to the other flagship models It is also very good in instruction following and agentic tool use benchmarks, again beating o3. OpenAI also showcased a bunch of other benchmarks, feel free to check those out in their introduction post. Context length for the model is 400k, which is roughly double the context size of the previous models. Max output tokens is 128k which basically means this is a beast for generating a lot of code. All in all the increased context size and output tokens basically mean that this should be great for MAX Mode in Cursor GPT-5 comes in 3 different variants that have different prices ($/Mtok): - GPT-5 - $1.25 in / $10.00 out - GPT-5-mini - $0.25 in / $2.00 out - GPT-5-nano - $0.05 in / $0.40 out The models are also already available in Cursor, both in the Cursor IDE as well as for Background Agents.