SpaceX AI and Cursor have released Grok 4.5, a jointly trained model that matches the reasoning capability of Claude Opus (Anthropic's flagship model) while costing 6× less per token. The model is already available in Cursor with expanded usage limits and through the Cursor API, offering developers a cost-effective alternative to premium closed models for production applications.
Summaries like this, in your inbox every morning.
Sign up free →What happened
SpaceX AI and Cursor jointly trained Grok 4.5, an Opus-class model that performs roughly between Opus 4.7 and 4.8. It is now available in Cursor with increased limits and via API. The model is 6× cheaper than Opus models and 3× cheaper than GPT-5.5 on a per-token basis while being token-efficient.
Why it matters
Developers and businesses using Cursor can now access frontier-class reasoning without the per-token cost of Claude Opus or GPT-5.5, reducing AI infrastructure spending for production workloads. This may reshape model selection decisions for cost-sensitive teams building applications.
What to watch
Grok 4.5 is available immediately in Cursor and via API. Early testers report it delivers comparable performance to higher-cost models, making token efficiency and pricing the key competitive factors in a crowded frontier-model market.
Grok 4.5 enters a market where cost per token has become a primary competitive lever alongside raw reasoning capability. The body reports that it achieves Opus-class performance while dramatically reducing inference costs, suggesting a shift toward efficiency-focused model selection. SpaceX AI and Cursor's joint training and release through Cursor's IDE indicate a deepening partnership to embed high-capability AI directly into developer workflows, competing on both capability and accessibility rather than just model rankings.
The framing in the body places Grok 4.5 alongside other recent model launches (GPT Live-1, Muse, SWE-1.7) that compete on specialized performance, cost, or speed. The emphasis on token efficiency and per-token pricing reflects a market where enterprises are optimizing not just for model quality but for total cost of ownership—a shift that may pressure developers to re-evaluate their default model choices even if those models rank higher on benchmarks.
No comments yet. Be the first to share your thoughts!
Log in to join the discussion
Get curated AI news from 200+ sources delivered daily to your inbox. Free to use.
Get Started FreeFree · takes 30 seconds · unsubscribe anytime
1 minute a day. The AI essentials.
200+ sources · Email / LINE / Slack
