Ring-1T — benchmark results
InclusionAI's open 1T-parameter Ring 1T reasoning MoE (October 2025). Provider: InclusionAI. Released 2025-07-01. Access: Open.
Unified ELO 1615 ± 24, rank #444 of 1776 rated models, from 68 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| AA AIME 2025 | 89.33 | Accuracy (%) | 88.5 |
| ConStory-Bench | 0.54 | CED errors per 10K words (lower is better) | 81.2 |
| UGI - Natural Intelligence | 33.69 | NatInt Score | 81 |
| Kagi LLM Benchmark | 66.4 | Accuracy (%) | 75 |
| OpenCompass Language - Dialogue | 98 | Score (%) | 74 |
| UGI Leaderboard | 41.69 | UGI Score | 71.4 |
| AA LiveCodeBench | 64.34 | Pass@1 (%) | 71.1 |
| CritPt | 0.6 | Accuracy (self-reported) | 71.1 |
| AA MMLU-Pro | 80.59 | Accuracy (%) | 69.5 |
| LLM2014 Logic 2025-10 | 46.07 | Median Score | 69.4 |
| AA GPQA Diamond | 77.37 | Accuracy (%) | 67.7 |
| LLM2014 Logic 2025-11 | 43.08 | Median Score | 65.4 |
Interactive version: theaggregate.ai/model?slug=ring-1t · How the rankings work · Data refreshed daily, snapshot 2026-07-22.