Model test results
Overall
Compare results in the same LiveBench suite. The first column is the equal-weight mean of its categories; each category is also shown separately. These are benchmark scores, not a universal rating.
Sort by test
LiveBench · Test suite version Oct 14, 2026Last checked 49 tested settingsClick a column heading to sort.
| # | Model and setting | ||||||||
|---|---|---|---|---|---|---|---|---|---|
| 1 | 88.0 | 80.8 | 91.7 | 97.7 | 100.0 | 80.3 | 82.0 | 83.4 | |
| 2 | 86.5 | 85.0 | 92.2 | 96.3 | 80.0 | 80.3 | 85.0 | 86.9 | |
| 3 | 86.1 | 82.4 | 92.7 | 93.7 | 84.0 | 83.0 | 85.2 | 82.0 | |
| 4 | 85.4 | 73.0 | 92.6 | 96.0 | 88.0 | 82.7 | 83.5 | 82.1 | |
| 5 | 84.1 | 69.9 | 95.1 | 89.7 | 92.0 | 85.2 | 83.1 | 73.7 | |
| 6 | 83.6 | 53.2 | 89.7 | 98.3 | 92.0 | 80.5 | 86.9 | 84.7 | |
| 7 | 82.5 | 78.8 | 91.6 | 94.7 | 84.0 | 59.5 | 83.6 | 85.0 | |
| 8 | 82.1 | 62.1 | 91.2 | 90.8 | 84.0 | 74.6 | 88.9 | 83.0 | |
| 9 | 81.9 | 66.9 | 88.7 | 93.1 | 92.0 | 81.2 | 69.3 | 82.2 | |
| 10 | 79.2 | 50.9 | 89.1 | 96.0 | 76.0 | 79.5 | 79.5 | 83.4 | |
| 11 | 78.1 | 53.2 | 90.6 | 86.2 | 72.0 | 81.6 | 88.0 | 75.1 | |
| 12 | 78.0 | 40.1 | 86.7 | 90.2 | 84.0 | 79.3 | 86.8 | 79.2 | |
| 13 | 76.8 | 48.3 | 89.2 | 77.3 | 96.0 | 66.0 | 81.8 | 78.8 | |
| 14 | 76.3 | 48.5 | 91.7 | 89.7 | 88.0 | 79.8 | 54.5 | 81.6 | |
| 15 | 75.6 | 57.6 | 82.4 | 91.4 | 80.0 | 53.1 | 81.9 | 83.1 | |
| 16 | 75.0 | 48.4 | 90.6 | 86.2 | 84.0 | 79.3 | 60.4 | 76.0 | |
| 17 | 74.9 | 40.7 | 91.2 | 89.1 | 64.0 | 75.3 | 84.5 | 79.8 | |
| 18 | 74.9 | 47.5 | 79.6 | 81.5 | 88.0 | 72.3 | 76.7 | 78.5 | |
| 19 | 74.8 | 40.4 | 89.7 | 90.2 | 80.0 | 79.6 | 68.6 | 74.8 | |
| 20 | 74.1 | 50.1 | 82.7 | 84.5 | 60.0 | 76.9 | 86.4 | 78.0 |
Scroll sideways to see all test columns.
What the columns measure
- Long agentic work:
- Long-running coding and app work across many turns
- Reasoning:
- Logic puzzles and problem solving
- Coding:
- Writing and completing code
- Browser use:
- Multi-step tasks in real web apps
- Data analysis:
- Questions about tables and data
- Research:
- Open-ended research reports
- Design:
- Layout and design of documents and slides
Tests and methodology
One test source: LiveBench. Columns are parts of this source, not independent benchmarks. Scores use the original 0–100 scale. Choose the column used for placement; a dash means missing data.
Each exact model setting stays separate. A category mean requires every task in that category; the overall mean requires every category. Tasks have equal weight within a category and categories have equal weight overall. Missing results are never zero. Placements use unrounded values in the chosen column, with shared ties; search retains the placement in the full test. Old scores cannot determine placement. Displayed decimals do not establish a significant difference. This view covers one test suite, not every current model or every kind of ability.
The suite date identifies the test version. LiveBench continues to add models and update results within that version; the check date shows when we fetched its live data.
LiveBench · Source ↗More tests and historical results →