Sphinx AI

Sphinx 1.0 is out now! Try it now at app.prod.sphinx.ai!

The best coding agents don’t understand data well enough to have good analytical instincts. Claude 4.5 Opus (right) confidently fits a single line to data with clear structure. The code runs, but the result is shoddy. Sphinx (left) is a real data scientist, not just a coder 👇

Read the full changelog here! sphinx.ai/blog/sphinx-0-…

Behind the scenes, we've been overhauling our agentic evaluation and alignment procedures, and users can expect many improvements to Sphinx's behavior, ranging from data science expertise to Jupyter notebook fluency. We'll be sharing more details on our research results soon!

Sphinx 0.9 is out now, featuring several long-requested product improvements, as well as the first agentic results from our recently overhauled research efforts. Read more below!

It's unfortunate to see this happen -- the point of climbing benchmarks is to help firms like ours deliver a better experience to users, not just to post fake high scores See more on how Sphinx is working through the leak here: sphinx.ai/blog/goodbye-d…

Thank you to @Adyen and @huggingface for the DABstep benchmark. This work helps define what great looks like and raises the bar for everyone. While Sphinx is moving away from DABstep as a primary benchmark after a test-set leak, we’ll still use it in a limited capacity. [1/2]

Lastly, we've doubled usage limits for all tiers of Sphinx, so that all users can experience using Sphinx more. Check out the full changelog here! sphinx.ai/blog/sphinx-0-…

Sphinx also is now able to perform Jupyter-aware file search. This is important because if your Jupyter server is hosted on a remote machine (like via Colab), your kernel will only be able to access that remote machine's file system. Sphinx understands this.

0.8 is now LIVE and it's the largest update yet, featuring Plan Mode, new Anthropic-powered engines, Jupyter-aware file search, and MORE!

As we head into the long weekend, I’m excited to dive into a few pet data projects. What’s making this ridiculously easy is Sphinx AI + the new Google Colab extension. In <1 min, I can spin up free GPUs for our best-in-class copilot. See how below! Added bonus: it's all free 💸

No compute? No problem! You can now use Sphinx with Google Colab's VS Code extension to build complex models and analyses in your IDE, all totally free and with no need for local resources! See Sphinx build a CNN model on MNIST on a Colab GPU:

Sphinx is now even faster! And, we've improved the knobs for how autonomously Sphinx acts as it works alongside you. sphinx.ai/blog/sphinx-0-…

See the full deep dive and more details on our blog! sphinx.ai/blog/not-all-d…

Even over a wide range of random scatterplots, frontier models only have a crude understanding of data! This is why Sphinx’s representation learning is critical to ensuring that data science gets done correctly.

Not all data agents are created equal. LLMs are fundamentally bad at understanding data, and without a good representation of data they make mistakes that are obvious to a human. This is what separates Sphinx from solutions where AI is bolted on to a notebook (1/N)