About Traictory
Traictory is an independent AI model intelligence platform. We track benchmarks, API pricing, and capabilities for every major language model — so you can make informed decisions without spending hours on research.
What We Do
The AI landscape moves fast. New models drop every week, benchmarks shift, pricing changes overnight. Traictory cuts through the noise by aggregating verified data from official API providers and independent evaluation papers into a single, searchable platform.
Whether you're a developer choosing between Claude and GPT for a coding project, a researcher comparing reasoning capabilities, or a startup evaluating API costs — our comparison tables, rankings, and analysis give you the data you need.
Our Methodology
Every model on Traictory is scored using a composite intelligence index based on 6 key benchmarks: GPQA (scientific reasoning), SWE-Bench (real-world coding), Tau2 (tool use), MMLU (broad knowledge), MMMU-Pro (multimodal), and ARC-AGI (abstract reasoning). We chose these benchmarks because they best differentiate frontier models — easier tests like GSM8K or HellaSwag have ceiling effects where most top models score 95%+.
Full methodology details, weight rationale, and academic citations are available on our methodology page.
Data Sources
Benchmark scores are sourced from official provider publications, peer-reviewed papers, and community-run leaderboards. Key academic references include GPQA (Rein et al., 2023), SWE-Bench (Jimenez et al., 2023), and MMLU (Hendrycks et al., 2020). API pricing is verified against official provider documentation and updated daily.
Independence
Traictory has no sponsorships, paid placements, or affiliate relationships with any AI provider. Our rankings are determined entirely by benchmark data. We believe independent evaluation is essential for a healthy AI ecosystem.
How We Work
Traictory uses AI-assisted tools for data aggregation, research, and content drafting — with all content reviewed and verified by our editorial team before publication. Every benchmark score, pricing data point, and model specification is traced back to an official source. We cite every reference so you can cross-check against official documentation for production decisions.
Who's Behind Traictory
AI technology researcher and founder of Traictory. Tracking model releases and benchmark results across 12 evaluation frameworks.
Contact
For questions about our data, methodology corrections, or partnership inquiries, reach out at team@traictory.com.