released 2026-07-151 source
Inkling benchmark scores: 16 benchmarks tracked. 31% of its scores are independently reproduced — the rest are self-reported or unverified.
Leads on 0 of 16 benchmarks·best result 88.88 on OTIS Mock AIME 2024-2025 (reproduced)·5 of 16 independently reproduced
| Benchmark | Score | Evidence status | Measured |
|---|---|---|---|
| OTIS Mock AIME 2024-2025 | 88.88accuracy (%) | reproduced· optimizedT1 | 2026-07-15Epoch AI |
| GPQA diamond | 84.34accuracy (%) | reproduced· optimizedT1 | 2026-07-15Epoch AI |
| ARC-AGI | 79.5accuracy (%) | unverified· optimizedT1 | 2026-07-15https://arcprize.org/leaderboard |
| DTBench | 79.12accuracy (%) | unverifiedT1 | 2026-07-15https://conceptualreasoning.ai/dtbench |
| LMCA | 44.21accuracy (%) | unverifiedT1 | 2026-07-15https://conceptualreasoning.ai/lmca |
| SimpleQA Verified | 40.3accuracy (%) | unverifiedT2 | 2026-07-15 |
| SimpleBench | 40accuracy (%) | unverifiedT1 | 2026-07-15https://lmcouncil.ai/benchmarks |
| ARC-AGI-2 | 36.53accuracy (%) | unverifiedT2 | 2026-07-15 |
| APEX-Agents | 33.8accuracy (%) | unverifiedT2 | 2026-07-15 |
| FrontierMath-Tiers-1-3-v2-Private | 33.33accuracy (%) | reproducedT1 | 2026-07-15Epoch AI |
| WeirdML | 32.29accuracy (%) | unverifiedT1 | 2026-07-15https://htihle.github.io/weirdml.html |
| Chess Puzzles | 16.88accuracy (%) | reproducedT1 | 2026-07-15Epoch AI |
| FrontierCode | 14accuracy (%) | unverifiedT1 | 2026-07-15https://cognition.com/frontiercode |
| FrontierMath-Tier-4-v2-Private | 4.88accuracy (%) | reproducedT1 | 2026-07-15Epoch AI |
| FrontierSWE | 4.12accuracy (%) | unverifiedT1 | 2026-07-15https://www.frontierswe.com/ |
| ProofBench | 0accuracy (%) | unverifiedT1 | 2026-07-15https://www.vals.ai/benchmarks/proof_bench |
Claims drawn from cited facts, not live model generation.
That recent vintage puts it at the leading edge of contemporary training and tooling practice, though with no stated parameter count its size class — compact, mid-size, or large — cannot be inferred.
1 cited fact
This model is scored on 16 tracked benchmarks. It holds no current top score on any of them. Its record can be trusted because the scores have been independently reproduced through outside evaluation.
3 cited facts
Its mean gap to the best recorded score across its tracked benchmarks is 46.98 points.
1 cited fact
1 facts cross-checked across data sources: 1 single-source.
Source facts, citations, and refresh stamp for this record.
Sources: epoch_benchmarks
Inkling is an AI model released 2026-07-15. Its benchmark record below tags every score as independently reproduced, self-reported, or unverified.
Inkling has recorded scores on 16 benchmarks, each shown with its evidence status.
Inkling has recorded scores on 16 benchmarks — GPQA diamond, OTIS Mock AIME 2024-2025, DTBench, WeirdML, LMCA, SimpleQA Verified, and 10 more. The full table above shows each score with its evidence status.
5 of 16 recorded scores (31%) are independently reproduced rather than self-reported by the lab.
Inkling has 16 tracked claims: 5 independently reproduced, 11 unverified.