Artificial Analysis's Long Context Reasoning metric as reported in the LLM Benchmark Matrix. Measures ability to reason over very long input contexts.
This definition has not completed editorial and provenance review. Verify its metric, version, source, and evaluation configuration upstream before using it for a model-selection decision.
Specifically targets long-context reasoning, not just retrieval (needle-in-haystack). Composite from AA.
Composite methodology is proprietary. Matrix snapshot may lag AA's own data.
Higher is better. Typical range 67–76 for frontier models.
Composite methodology is proprietary.
| # | Model | Score |
|---|---|---|
| 1 | 91.6 | |
| 2 | 90.1 | |
| 3 | 90.1 | |
| 4 | 89.8 | |
| 5 | 89.1 | |
| 6 | 88.1 | |
| 7 | 88.0 | |
| 8 | 87.9 | |
| 9 | 87.8 | |
| 10 | 87.2 | |
| 11 | 86.4 | |
| 12 | 85.2 | |
| 13 | 81.5 | |
| 14 | 80.2 | |
| 15 | 80.2 | |
| 16 | 79.8 | |
| 17 | 79.4 | |
| 18 | 79.4 | |
| 19 | 79.3 | |
| 20 | 79.1 | |
| 21 | 78.9 | |
| 22 | 78.8 | |
| 23 | 78.7 | |
| 24 | 78.6 | |
| 25 | 78.6 | |
| 26 | 78.6 | |
| 27 | 78.4 | |
| 28 | 78.3 | |
| 29 | 78.3 | |
| 30 | 77.9 | |
| 31 | 77.9 | |
| 32 | 77.9 | |
| 33 | 77.8 | |
| 34 | 77.8 | |
| 35 | 77.8 | |
| 36 | 77.7 | |
| 37 | 77.5 | |
| 38 | 77.3 | |
| 39 | 77.3 | |
| 40 | 77.3 | |
| 41 | 77.3 | |
| 42 | 77.3 | |
| 43 | 77.2 | |
| 44 | 77.2 | |
| 45 | 77.2 | |
| 46 | 77.2 | |
| 47 | 77.1 | |
| 48 | 77.1 | |
| 49 | 77.0 | |
| 50 | 77.0 |