--- title: "Artificial Analysis Ships Model Sets to Cut Repetitive Comparisons Across 542 Models" subtitle: "Artificial Analysis launches Model Sets, letting users save and instantly apply custom model groups across all benchmark charts on the platform." url: "https://alphasignal.ai/news/artificial-analysis-ships-model-sets-to-cut-repetitive-comparisons-across-542" author: "AlphaSignal Newsroom" author_url: "https://alphasignal.ai/editorial-team" company: "Artificial Analysis" company_url: "https://alphasignal.ai/company/artificialanlys" published: "2026-06-26T19:21:10.000Z" updated: "2026-06-26T21:32:06.381Z" type: "News" topics: - "Benchmarks" - "Llms" subtopics: - "Hallucinations" - "Long Context" - "Small Models" upvotes: 304 read_time: "3 min" source_url: "https://x.com/artificialanlys/status/2070588196975784339" comments: 0 --- # Artificial Analysis Ships Model Sets to Cut Repetitive Comparisons Across 542 Models > Artificial Analysis launches Model Sets, letting users save and instantly apply custom model groups across all benchmark charts on the platform. [Artificial Analysis](https://artificialanalysis.ai) just shipped a quality-of-life feature that its users have been asking for: **Model Sets**. You can now save a named collection of models and apply it across every chart on the platform in one click, instead of re-selecting the same handful of models every time you open a new view. ## The problem it solves If you have spent any time on Artificial Analysis, you know the drill. The platform lets you create custom visualizations, compare models and providers, and export data across a wide range of dimensions. Trade-offs exist between model quality, price, output speed, latency, context window, and other dimensions, and most people are not comparing all 542 models at once. They have a shortlist. The old workflow meant rebuilding that shortlist from scratch on every chart. If you were tracking, say, five frontier models across the Intelligence Index, the coding leaderboard, the speed chart, and the provider comparison, you were clicking through the same model selector four separate times. Model Sets collapses that into one saved action. ## How it works The feature lives inside the [Data Playground](https://artificialanalysis.ai/data-playground), where users can create custom visualizations and tables comparing models and providers, save groups of models, and export data. Once you define a set, it becomes a persistent selection you can apply globally. The workflow looks roughly like this: 1. Open the model selector on any chart 2. Pick the models you care about 3. Save the selection as a named Model Set 4. On any other chart, load that set instantly from the selector The feature is live now and available to all users on the platform. ## Why this matters for day-to-day use Artificial Analysis tracks a lot. The platform maintains a ranked list of over 100 LLMs based on multiple metrics: Intelligence Index, speed, latency, price, and context window. The model count on the full leaderboard is now over 542. Navigating that without a saved filter is friction. The people who feel this most are the ones doing repeated, structured comparisons: - **Researchers** tracking a fixed set of frontier models across new benchmark releases - **Engineering teams** evaluating a shortlist of candidates for a production deployment - **Procurement teams** monitoring cost and speed trends for models already in use - Consultancies, enterprise platforms, or procurement teams who need benchmarking as a service The platform's visualization tools display trade-offs, such as how a model's intelligence score correlates with its latency, helping stakeholders balance cost and capability. Model Sets make it much easier to keep that analysis consistent across sessions and across colleagues. ## The broader platform context This is a UX addition on top of a platform that has been expanding its evaluation surface aggressively. The Artificial Analysis Intelligence Index v4.1 incorporates 9 evaluations: GDPval-AA v2, τ³-Banking, Terminal-Bench v2.1, SciCode, Humanity's Last Exam, GPQA Diamond, CritPt, AA-Omniscience, and AA-LCR. The team has also recently shipped AA-Briefcase, a new agentic benchmark for long-horizon knowledge work, and a Speech-to-Speech Index. Since its launch in 2023, Artificial Analysis has seen significant growth in user engagement, with website traffic increasing by 27.56% month-over-month as of late 2025. As the model count balloons and the benchmark suite deepens, features like Model Sets are the kind of infrastructure that keeps the platform usable at scale. Model Sets is available now at no additional cost. The Data Playground, where the feature is anchored, is part of the [premium plan tier](https://artificialanalysis.ai/pricing), though the core leaderboard charts are free. If you are already a regular user of the platform's comparison views, this is a straightforward upgrade to your workflow. ## Sources - [X post by @artificialanlys](https://x.com/artificialanlys/status/2070588196975784339) - [Artificial Analysis – Official Homepage](https://artificialanalysis.ai/) - [Artificial Analysis – Model Comparison](https://artificialanalysis.ai/models) - [ArtificialAnalysis.ai: The Ultimate Guide for Developers and Business Leaders | Medium](https://cgorale111.medium.com/artificialanalysis-ai-d62d5e3a63e4) - [Artificial Analysis – Guide](https://artificialanalysis.ai/guide) - [Artificial Analysis – Data API](https://artificialanalysis.ai/data-api)