TASK 1 / TEST RESULTS

Test results

27 teams · 59 accepted submissions · 928 questions. Published 28 October 2026 using the frozen evaluation configuration.

EVALUATION NOTE

Known question inconsistency

One evaluation question has a confirmed inconsistency between its wording and frozen reference answer. These published scores retain the original frozen evaluation; no correction or exclusion has been applied. Any later adjudication will be announced with a versioned results update.

Test · frozen scorer 2.0.0

Rankings

27 teams

Rank	Team	Combined FAC ↓	Hidden FAC	Combined checkpoint	Final answers correct / 928
1	SJX	0.999167	1.000000	0.999583	927
2	sader	0.999167	1.000000	0.997500	927
2	FinVeritas	0.999167	1.000000	0.997500	927
4	Patodiya Research	0.999167	1.000000	0.995417	927
5	nitishxbt	0.998175	0.998016	0.996111	926
6	NIA.XS	0.997500	1.000000	0.996111	925
7	EDB	0.997500	1.000000	0.995417	925
7	DSsama	0.997500	1.000000	0.995417	925
9	bluck	0.997341	0.998016	0.994286	925
10	NTaiwanUs	0.996458	1.000000	0.995590	924
11	HumanSlop	0.995466	0.998016	0.992614	923
12	Infinity Analyze	0.994077	0.995238	0.992614	922
13	Ledger Logic	0.994008	0.989683	0.960810	923
14	ni shuo de dou dui	0.992341	0.989683	0.953155	921
15	VeriFin	0.991429	0.984524	0.993631	920
16	SSB	0.990308	0.987698	0.985967	919
17	carnot	0.985304	0.976025	0.988745	917
18	xai	0.982887	0.974107	0.975109	913
19	TieOut	0.980546	0.969841	0.976920	912
20	KevinSoLab	0.973929	0.994444	0.983108	896
21	besk.tech	0.971042	0.943750	0.983016	906
22	nailong	0.968373	0.942460	0.969494	908
23	Push and Pray	0.967870	0.943122	0.974577	904
24	Metric Max	0.961224	0.929828	0.952209	900
25	ABBO	0.916862	0.838724	0.937222	872
26	Math	0.488333	0.000000	0.489861	566
27	ut1776	0.336389	0.055456	0.345046	377

Ranked by Combined FAC, Hidden FAC, then Combined checkpoint, descending after six-decimal round-half-even quantization. Identical three-key scores share a competition rank; submission time is not a tie-breaker.

Each team is scored on its last accepted, valid submission before 25 October 2026, 23:59 AoE (26 October, 11:59:59 UTC), not its best upload. One organizer test is excluded. ABBO and ji share one backend identity and are ranked once under the final name ABBO.

Combined FAC and Combined checkpoint each give 50% weight to the seen and hidden Test strata, using the published macro-averaging rules. They are not Dev/Test averages or simple case-count accuracies. Test answers remain private. Organizer baseline Test scores are not included.

Download results CSV · Download results JSON

Submission receipts

Competition submissions are closed. The original portal remains available for submission-status information. During the competition, a format pass alone was not an acceptance receipt; no test score, rank, or answer-correctness feedback was returned.

Open submission portal

Working Notes are due 2 November 2026, 23:59 AoE. Paper instructions · Development leaderboard