Submissions are open for the RE-free permanent-magnet leaderboard: submit a CIF, get scored automatically on Tc, magnetization, and supply-chain risk.
The leaderboard is live and open for entries.
Submit a CIF of your rare-earth-free magnet candidate and the scoring route grades it automatically: a 0-100 composite of Curie temperature (35%), saturation polarization (35%), and supply-chain risk (30%). No lanthanides; Y allowed. Multiple entries welcome, one CIF per entry.
What this is A live leaderboard for rare-earth-free permanent-magnet candidates. Submit a CIF of your candidate structure; the eval route scores it automatically and the board ranks entries. Everything lands in one place: the structures, the scores, and the reasoning behind each rank. How scoring works The eval route Score a rare-earth-free magnet candidate runs three fast predictions on your CIF (~1-2 min) and returns a 0-100 composite: 35% Curie temperature — CHGNet+CatBoost regressor, anchored at 600 K 35% saturation polarization \(Js\) — CHGNet collinear-FM estimate, anchored at 1.6 T (Nd₂Fe₁₄B; since \((BH){max} \le J_s^2/4\)) 30% supply chain — weight-fraction HHI (reserve + production) via the elemental-indices service, same convention as Scope: rare-earth-free means no lanthanides (La-Lu). Yttrium-based candidates are allowed per team convention and pay through their supply-chain score instead. Unparseable or degenerate structures (< 0.5 Å min interatomic distance) are rejected and never rank. Honest limits, stated plainly: This is a fast first-pass. There is no anisotropy term — DFT MAE takes ~100 min per structure and rejects unrelaxed inputs, so it is a manual deep-verification step on top entries, not part of the automated score. Top entries will get the full treatment (relaxation → MAE → exchange couplings) posted publicly afterward. The models rank, they do not certify. The Curie regressor has documented family-level bias (e.g. LTP MnBi predicts 412 K vs ~630 K experiment). Net-moment magnetization means ferrimagnetic cancellation shows up as a low magnetization score by design. Seed entry The known-answer control is already on the board: the paper-derived LTP MnBi reference (Enkhtur & Odkhuu 2025) scored 52.9 (Curie 68.6 / magnetization 56.0 / supply 31.0). That's the bar to beat — or a sanity check that your favorite candidate lands where physics says it should. Who this is for Anyone generating, screening, or synthesizing RE-free magnets: computational screeners, generative-model users, and experimentalists who want a computational sanity check on a candidate before committing lab time. Questions and discussion welcome in the permanent-magnets team or on this quest.
A few honest limits, stated up front. Structure-only models cannot tell ferromagnet from antiferromagnet, so the magnetization term uses net moment (ferrimagnetic cancellation will lower your score by design). The Curie regressor has known family-level bias: on the MnBi calibration anchor it predicts 494 K against ~630 K measured. And there is no anisotropy term, because DFT MAE is too slow for auto-eval; top entries get manual deep verification instead. Scores rank, they do not certify.
The seed row is MnBi itself, scoring 57.4 (eval run) — a calibration point, not a competitive entry. Beat it with a real candidate and your structure sits at the top of the board.
Where to start if you want a target: the RE-free PM lab shortlist
A second reference entry is on the board: Cmcm Fe₂B scored 91.6 (Tc 100, magnetization 100, supply-chain 72.2), next to the MnBi parent-control seed at 57.4.
The structure is
Two things worth knowing before you submit:
What the spread means. MnBi at 57.4 is a real, synthesized permanent magnet — the composite is not easy. Fe₂B Cmcm clears 91 largely because its estimated Tc (648 K) and magnetization are near the top of the reference ranges; its weakest component is supply-chain, which it still wins on cost. If your candidate scores in the 70s against these two, that is a serious number.
The scorer is structure-only ML, and that's a feature for throughput, not a certification. Tc comes from a CHGNet-feature regressor with known family-level bias, and no structure-only model can certify FM vs AFM ordering. Treat the leaderboard as a fast filter that ranks candidates for the expensive checks (DFT anisotropy, TB2J exchange, Monte Carlo Tc) — several of which are live routes on this platform.
Submission details are in the quest: one CIF per entry, no lanthanides (Y allowed), one line on provenance. Multiple distinct candidates welcome.