THE AI LANDSCAPE 08 / MODEL BENCHMARKS
Compare AI models by intelligence score, explore each developer’s lineup, and follow the evidence. A vertical leaderboard with company filters, reasoning configurations, and dated sources.
THE INTELLIGENCE INDEX
One model configuration per row. Compare scores, explore a company, and open the original analysis.
Try a different developer or model name.
Scores: Artificial Analysis · v4.3.2 · Higher is better
This dated snapshot includes all 273 entries in Artificial Analysis’s Current-status LLM leaderboard on October 21, 2026: 268 scored configurations and five without an index score, across 57 developers. Reasoning settings are separate results. This covers models listed by that source, not every product a developer has released; image, video and speech-only models are outside this index. The snapshot is updated editorially, not live.
“Models” shows the highest-scoring configuration for each model name within a developer’s lineup. “All configurations” includes every reasoning setting and other variant listed in the source. Company and search filters apply to the selected view. Rank numbers refer to the full selected view, before filtering.
An asterisk (*) preserves a score caveat shown by Artificial Analysis. Follow the model link for its original analysis and qualifications. A dash means the source did not publish an index score; it is not a zero. Unscored entries appear after scored results in the default order.
Intelligence Index v4.3.2 combines ten evaluations spanning agent tasks, reasoning, knowledge and coding. It is primarily text-based and English-language. A higher index score does not measure price, speed or suitability for every task. Different reasoning settings can produce different results for the same model.
The source’s displayed whole-number scores are preserved. Equal scores share a rank; their display order is not a tie-break. All bars use the same 0–60 axis, even when filtered. Index points are not percentages. A benchmark listing does not guarantee public access to a model.
AI Rankings brings model intelligence scores into one clear, vertical leaderboard. Explore developers, compare model names and reasoning configurations, and follow each result to its original analysis.
The benchmark snapshot was checked on October 21, 2026. It includes 273 configurations from 57 developers in the Current-status Artificial Analysis LLM leaderboard. The default Models view shows 192 model names, using each model’s highest-scoring listed configuration. The All configurations view includes every listed result, including entries without a published score.
Search for a model or developer, select a company, or change the order to compare lineups. Switch between Models and All configurations to explore reasoning settings. Show more results or download every matching result as a CSV with source links.
The leaderboard uses the Artificial Analysis Intelligence Index v4.3.2. Higher scores indicate stronger performance on that index; they are points, not percentages. All bars use the same 0–60 scale. Equal displayed scores share a rank, and a dash indicates that no score was published. Asterisks preserve qualifications in the original source.
Benchmark performance is one part of choosing a model. Price, speed, availability, licensing and suitability for your workload also matter. This is a dated editorial snapshot of the models covered by the source, not a live feed or a complete inventory of every developer’s products. Image, video and speech-only models are outside this index.
See the Artificial Analysis leaderboard and its benchmarking methodology for the original results and evaluation details.
Use AI Models to explore selected developer lineups, model families, access and open weights. Visit AI Tools to find applications, or return to AI Landscape for more maps and comparisons.