05November 2026 · Research
OpenAI's 721-paper mathematics corpus as one Flywheel graph, and four follow-up results found by agents working on it, written up in full below.
We ported OpenAI's 721-paper mathematics corpus to Flywheel as one graph of 7,095 linked statements and ran agents on it. Four of the follow-up results are written up here: an explicit sensitivity exponent, an Artin group on 24 generators, a larger lower bound for two-way automata, and a Navier–Stokes proposition that claims more than its proof gives.
Muhammed Emin Baslak, Giulio Starace, and Francesco Pappone
04September 2026 · Research
Paradigma's first model, built for high-throughput mathematical reasoning.
A 1-billion parameter transformer trained from scratch for mathematical reasoning, rivalling models tens of times larger on competition-level math. Released with its evaluations, a training-time value model, and a custom vLLM inference plugin.
Mario Prignano, Gabriele Cirillo, Alessio Morosini, Luca Cerovaz, Alessandro Bartolocci,
Emanuele Rodolà, Giulio Starace, and Francesco Pappone
03September 2026 · Blog
How results could be shared in the new age of research, using the Riemann Hypothesis as a working example.
As models get better, everyone is tempted to spend their compute on the big open problems. We are still figuring out how to collect and share that work so others can build on it. Here is a prototype of what we envision, with a result on Gröchenig's Riemann hypothesis criterion shared as a Flywheel graph.
Paradigma
02August 2026 · Research
How the Via Panisperna Fellows set records across four frontier AI training challenges.
The ideas behind our work on the Modded-NanoGPT Main and Optimization Tracks, OpenAI's Parameter Golf, and the SlowRun Tiny Track—and what the process taught us about collaborating with coding agents.
Luca Cerovaz, Mario Prignano, Alessandro Bartolocci, Gabriele Cirillo, and Donato Crisostomi
01March 2026 · Blog
Pointing intelligence at the edge of the unknown.
Every hard problem humanity has not solved yet is, at its core, a problem of insufficient intelligence applied for insufficient time. This essay argues that if intelligence becomes scalable, the infrastructure that organizes its output becomes load-bearing.
Francesco Pappone