Surge AI

A model post-trained on non-coding office tasks gained 5.7 points on SWE-Bench Pro. Why? Our analysis points to a capability we call Goal-Directed Execution.

Measuring AI model performance on extreme-tier mathematical problems.

Can frontier models master the documents that run the world? GDP.pdf is a professional multimodal reasoning benchmark that takes real-world prompts and PDFs pulled directly from expert enterprise workflows.

Stop testing models in tiny, self-contained environments. We built CoreCraft, a large-scale startup world, and deployed AI agents to solve real tasks. Our goal: to move agents beyond the cleanliness of the lab and into the chaos of enterprise reality.

Stop rewarding slop. Hemingway-bench is an AI writing leaderboard that takes real-world writing tasks and puts them in front of master wordsmiths. Our goal: to push AI writing from two-second vibes to genuine nuance and impact.

Surge AI CEO Edwin Chen said data labeling is often over-simplified. In his view, the work is "a lot more like raising a child."

Meta Superintelligence Labs partnered with Surge to build AdvancedIF, an instruction-following benchmark where every prompt and rubric was written by human experts - not synthetically generated by an LLM. In instruction-following domains, where frontier m

Would you trust a medical system whose only metric was

Would you trust a medical system whose only metric was

We stress-tested GPT-5, Gemini 2.5 Pro, and Claude Sonnet 4.5 on 200+ expert finance tasks. Here's where even the best models break when they move from benchmarks to Wall Street.