Knowledge · Benchmark profile
A display-only MMLU-Pro reference from Arcee AI's Trinity-Large-Thinking launch chart.
Data verified 21 Jul 2026 · Methodology 1.6.0
Benchmark score on MMLU-Pro first-party comparison snapshot
Claude Opus 4.6 leads at 89.1%, followed by Kimi K2.5 (87.1%) and GLM-5 (85.8%).
Visual analysis
Switch between model placement, score distribution and descriptive provider averages. Every view uses the same sourced leaderboard.
6 ranked models · higher is better
| Model | Rank | Provider | Score |
|---|---|---|---|
| Claude Opus 4.6 | #1 | Anthropic | 89.1% |
| Kimi K2.5 | #2 | Moonshot AI | 87.1% |
| GLM-5 | #3 | Z.AI | 85.8% |
| Trinity-Large-Thinking | #4 | Arcee AI | 83.4% |
| MiniMax M2.7 | #5 | MiniMax | 80.8% |
| Trinity-Large-Preview | #6 | Arcee AI | 75.2% |
One best score per model · higher is better
| Rank | Model | Provider | License | Evidence use | Score |
|---|---|---|---|---|---|
| #1 | Claude Opus 4.6 claude-opus-4-6 | Anthropic | closed | Reference only | 89.1% |
| #2 | Kimi K2.5 kimi-k2.5 | Moonshot AI | — | Reference only | 87.1% |
| #3 | GLM-5 glm-5 | Z.AI | open | Reference only | 85.8% |
| #4 | Trinity-Large-Thinking trinity-large-thinking | Arcee AI | open | Reference only | 83.4% |
| #5 | MiniMax M2.7 minimax-m2-7 | MiniMax | open | Reference only | 80.8% |
| #6 | Trinity-Large-Preview trinity-large-preview | Arcee AI | open | Estimated reference | 75.2% |
The top of this snapshot is led by Claude Opus 4.6 at 89.1%; third place is 3.3 points behind. The top-6 spread is 13.9 points.
About MMLU-Pro first-party comparison snapshot
A display-only MMLU-Pro reference from Arcee AI's Trinity-Large-Thinking launch chart. Results stay tied to the exact model variant and evaluation system. Multiple systems for the same model use the best published score on this page; overall Lumina scoring uses the median of ranking-eligible rows.
Open benchmark source ↗FAQ
A display-only MMLU-Pro reference from Arcee AI's Trinity-Large-Thinking launch chart.
Claude Opus 4.6 by Anthropic currently leads with 89.1%.
6 models in the LuminaBench cohort have a qualifying score on this benchmark.
No. This benchmark is display-only and does not enter the overall Lumina composite.
Related
2026 · 10 results · Reference
Artificial Analysis MMLU-Pro2026 · 4 results · Reference
Massive Multitask Language Understanding2020 · 8 results · Reference
Graduate-Level Google-Proof Q&A2023 · 70 results · Reference
SuperGPQA: Scaling LLM Evaluation Across 285 Graduate Disciplines2025 · 19 results · Reference
Massive Multitask Language Understanding Professional2024 · 42 results · Reference
