Ulam C Verified
RL environment company · Math
- On X
- @ulamdotai ↗
- Tier
- C why
- Snapshot
What Ulam does
Ulam sells research-grade mathematical reasoning datasets, private benchmarks and evaluations, and verifier-backed RL gyms for frontier AI labs.
- Sells
- RL environments and Data
- Leads with
- Math
- Based in
- Warsaw, Poland
- Team
- 1-10
- Website
- ulam.ai ↗
What Ulam sells
Read from Ulam's own website, checked October 12, 2026: the line that puts it in each category.
RL environments
“Custom benchmark suites and eval protocols for labs that need private, difficult-to-overfit measures of frontier reasoning.”
ulam.ai ↗“Train and evaluate agents in stateful, verifier-backed research environments where progress, revision, and failure become learning signals.”
ulam.ai ↗
Data
Domains Ulam covers
Math leads with
“Our work is grounded in open mathematical research, where progress is measured by the quality of reasoning, not just final answers.”
ulam.ai ↗“train and evaluate agents in verifier-backed RL gyms for code, algorithms, science, and mathematics.”
ulam.ai ↗“train and evaluate agents in verifier-backed RL gyms for code, algorithms, science, and mathematics.”
ulam.ai ↗
Ulam competitors and alternatives
The companies that sell the most of the same things in the same domains, and what they share with Ulam.
- ReasonCore Also sells datasets, RL environments, evals and human data · also in science, math and coding
- Revelo Also sells RL environments, evals, human data and datasets · also in coding and science
- Turing Also sells human data, RL environments and datasets · also in coding, science and math
- Vetto AI Also sells RL environments, evals, human data and datasets · also in science and coding
- Abundant Also sells RL environments, datasets and evals · also in coding and science
Tier C
Early: a small team.
How much stands behind a tier: usage, who backs it, a confirmed line from its own site or paper, and a clear read on X. High with three of the four, medium with two, low otherwise.
- team 1-10
- in curated RL directories (3 pts)
On X · mixed, low buzz, from 13 posts about Ulam in the last month
Founder Przemek Chojecki's math problem lists draw hostile replies over unverified statements and loose P=NP framing; STEMBench gets some praise as an interesting research benchmark.
How we got to tier C
- Adoption: team 1-10. Tier C.
- On X: mixed, low buzz. “Founder Przemek Chojecki's math problem lists draw hostile replies over unverified statements and loose P=NP framing; STEMBench gets some praise as an interesting research benchmark.” No move.
- In curated RL directories: 3 points, a floor of none.
- Confirmed: a line from its own site or paper places it.
A clear view on X moves a tier one step up or down; most entries don't move. How tiers work.
Training methods
Methods Ulam uses, offers or builds for, with public evidence: its own website says so, or two independent sites do.
- RLVR Reinforcement Learning with Verifiable Rewards On its own website
“2.564 / 4 A-SFT + V-RLVR · rank 2”
ulam.ai“Strict RLVR rows are an explicitly promoted subset—not a label for the whole corpus. The public release provides 10 inspection examples, schema, and quality-gate definitions.”
ulam.ai - DPO Direct Preference Optimization On its own website
“Training views: SFT, DPO, critic and repair data, RLVR/GRPO tasks, process-reward steps, private evaluation, and regression tests.”
ulam.ai“Preferences Verifier-ranked pairs Failed and corrected attempts become DPO-compatible signal.”
ulam.ai - GRPO Group Relative Policy Optimization On its own website
“Training views: SFT, DPO, critic and repair data, RLVR/GRPO tasks, process-reward steps, private evaluation, and regression tests.”
ulam.ai“Identity Exact checkpoint lineage Base, SFT, DPO, and GRPO students remain distinguishable.”
ulam.ai - SFT Supervised Fine-Tuning On its own website
“Tests shaped verification rewards alone and after specialized SFT, separating proof-critique skill from broad completion reliability.”
ulam.ai“Turns 5,387 TeX source files into structured proof-process data for long-context SFT, proof reconstruction, theorem dependencies, process supervision, critic training, preferences, and adversarial verifier tests.”
ulam.ai
Where Ulam sits
Its domain on the map, live. Hover a logo for its name; click to open it. Open it on the full map
Next: browse all environments. Entry from the RL Research daily scrape of public sources, 2026-11-05 snapshot.