Timeline & Prizes

Schedule & rewards

A three-month development phase, a blind final, and recognition that goes well beyond the cash awards.

Schedule

Key dates

  1. Feb 2026 Complete

    Dataset assembly and curation

  2. Mar 2026 Complete

    Baseline code, evaluation scripts, dFL visualizer

  3. Apr 2026 Complete

    Website online

  4. Jul 2026 Complete

    Public release of training & public test data

  5. Jul 27 – Oct 18, 2026

    Phase 1 (development) — continuous public leaderboard

  6. Oct 19, 2026

    Blind test data released; Phase 2 (final) opens

  7. Oct 26, 2026

    Phase 2 (final) closes — last call for submissions

  8. Nov 2026

    Verification, peer review of top entries, prize allocation

  9. Dec 2026

    NeurIPS competition session & results

  10. Q1 2027

    Lessons-learned community paper with selected winners

Cash awards

$1,000 — one $500 award per challenge

There are two challenges and two awards, matched one to one. You may enter either or both; entering both makes you eligible for both.

Award #1 · Challenge 1 $500

Best DIII-D intra-machine reconstruction

Highest composite S_model on the hidden DIII-D test set.

Award #2 · Challenge 2 $500

Best DIII-D → MAST cross-machine generalization

Highest G_ratio among entries reaching a DIII-D composite S_model ≥ 0.85 (gate enforced by the scorer). Needs predictions for both machines — a DIII-D-only entry scores G_ratio G_ratio The score that decides Challenge 2: G = S_model(MAST) / S_model(DIII-D) — the fraction of your DIII-D performance you keep when moving zero-shot to MAST. 1.0 would mean perfect transfer. It needs predictions for both machines, and only entries with S_model(DIII-D) ≥ 0.85 are eligible — below that gate (or with DIII-D missing) the scorer reports 0. = 0.

Live leaderboard Phase 1 — development

Mirrored from Codabench, which is the source of truth for every score. 25 entries · as of Aug 18, 2026

# Participant Ch1: DIII-D SCh2: G_ratio
1 pshishkin 0.9938 0.5202
2 dgaytanv 0.9932 0.0
3 dmitryzhevn 0.9932 0.7299
4 mnn31 0.9921 0.5642
5 catskills 0.9918 0.0558
Full board on Codabench → Showing the top 5 · ranking here is indicative; awards are decided on the Phase 2 (final) board.

Four honorable mentions

Top flux fidelity (R²ψ)
Top predicted-scalar fidelity (R²_q95,βN)
Top flux-map consistency (the ψ-derived scalars)
Best LCFS alignment (D_LCFS)
Beyond the cash

Recognition & support

Co-authorship & talks

Top three teams in each award category receive named co-authorship on the lessons-learned paper plus an invited podium talk at the NeurIPS competition session.

Certificates & talks

The four honorable-mention teams receive certificates and podium talks.

Compute for everyone

We are seeking sponsored GPU and cloud credits for participants from underrepresented or resource-constrained institutions. Nothing is confirmed yet — any that we secure will be posted on Announcements. No entry depends on them; see below.

Accessible by design

Both awards are reachable without dedicated hardware. The compute-light badge Compute-light badge A self-declared badge for models that train end-to-end in under two hours on a single GPU or CPU. It is a recognition that applies across both challenges — not a separate track, and not a separate leaderboard. celebrates entries that train in under two hours on a single commodity GPU or CPU — it applies across both challenges, and every reference baseline qualifies. See the challenge and the baselines for what that looks like in practice.