Grok 4 Fast is xAI's latest multimodal reasoning model, setting a new standard for cost-efficient intelligence. With a 2M token context window, it achieves a SOTA price-to-performance ratio. It's now available for free in the Grok app.
xAI's new Grok 4 Fast is built around a pretty compelling pitch.
It's fast (~190 tokens per second). It's cheap (API pricing is $0.20 / 1M input and $0.50 / 1M output tokens). And according to Artificial Analysis, this gives it a new SOTA price-to-intelligence ratio.
But the most important part is the accessibility. For a limited time, it's completely free for developers on platforms like OpenRouter. And it's also free for all users in the Grok app on web, iOS, and Android.
So, when a model is this fast, this cheap, and free for everyone to try right now... what's the reason not to?
Grok 4 Fast is xAI’s latest leap in multimodal reasoning 2M token context, SOTA price-to-performance, and now free in the Grok app. Built for speed, scale, and smarter decisions at lower cost.
“Grok 3 excels in reasoning, text, and code-related tasks, making it especially suitable for developers and users who need efficient text processing. However, there is still room for improvement in image generation and global accessibility. Its positioning is more like a ‘work-oriented AI assistant’ rather than a tool focused on entertainment or creativity.
I noticed your chat interface looks really clean and lightweight. Compared to products like ChatGPT or Claude, what do you see as your biggest differentiator?
Thanks for sharing. As a Grok user I am pleased. Trying it out... it seems to perform well.
The 2M token context window combined with that price-to-performance ratio is genuinely impressive. The fact that it's processing ~190 tokens/second while maintaining SOTA reasoning capabilities suggests we're seeing the beginning of a new paradigm where speed, cost, and intelligence no longer require trade-offs. Curious to see how this pressure forces competitors to respond on both pricing and architecture fronts.
This looks really cool! I'm really impressed by how Grok 4 Fast makes AI more accessible without sacrificing quality. That cost, efficient approach is a game changer for so many developers and businesses.
I'm curious, how does the fast processing feature work? Does it use specific algorithms to speed things up, or is it more about resource management?
I've seen so many discussions about how people are struggling to afford the high costs of AI solutions. Your product seems like it could really help out in that area.
If you're looking to connect with more developers or businesses facing these challenges, I know some groups that would definitely appreciate what you're building. Engaging with those discussions could really highlight how Grok 4 Fast can make a difference.
Grok 4 Fast looks amazing! I can already see it being super useful for long-form reasoning and complex prompts.
Congrats! What steps is xAI taking to address Grok 4 Fast’s extreme sensitivity to small prompt changes, which causes inconsistent outputs and makes it difficult to rely on the model for production workflows?
2M token context window is impressive, but curious about the memory optimization under the hood? Handling long sequences without OOM errors ain't easy.
Hi everyone!
xAI's new Grok 4 Fast is built around a pretty compelling pitch.
It's fast (~190 tokens per second). It's cheap (API pricing is $0.20 / 1M input and $0.50 / 1M output tokens). And according to Artificial Analysis, this gives it a new SOTA price-to-intelligence ratio.
But the most important part is the accessibility. For a limited time, it's completely free for developers on platforms like OpenRouter. And it's also free for all users in the Grok app on web, iOS, and Android.
So, when a model is this fast, this cheap, and free for everyone to try right now... what's the reason not to?