Claude Sonnet 4.5 pricing:
Starting at $3/M input tokens
Starting at $15/M output tokens
If you are using OpenRouter I can highly recommend this model to save you money. We have oneshotted a load of fixes (including some crazy complex issues). It is also incredibly fast!
OpenRouter is great as can select from all models not just those Copilot has to offer. Just add your OpenRouter key to Wappler's AI configuration and boom! Models like the new Minimax 2.1 are ground-breaking to say the least. It has solved some issues where Claude would just go off on one burning tokens without ever resolving the issue. Can afford to do this with Minimax 2.1 as the cost is so minuscule, literally a couple of Cents, meaning it can go all day long and not burn a hole in your wallet.
We have been using Minimax for quite a while now and the latest Minimax 2.7 is mind-blowing! Released yesterday its very close to Opus 4.6/Sonnet 4.6 and Codex 5.3... You can bash it all day long and barely notice a shift in your Open Router balance... We have it doing thousands of translations, creating blog posts autonomously, carrying out statistics and analytics, banging out reports and mailing them, and a ton of other functions. Its writing style is hard to separate from a human. It follows instructions absolutely and to the point... We have a crazy complex code base in one Project and it is working with it like it created it (Opus did a lot of the work). I said 2.1 was astonishing, and 2.5 even more so, and now 2.7 I'm quite lost for words... It is now training and improving itself. So I don't doubt it will be too long before Minimax 3.0 is available at the rate this lab is pushing boundaries!
M2.7 is really good, especially for the price when using APIs for it compared to Sonnet 4.6. Still a little uneasy with how much IP I give it, but it saves a ton of money for similar quality output.
