New Leaders Emerge in AI Benchmarking
Recent benchmark results reveal new top performers and first-time measurements in AI model evaluations.
Epoch AI has reported that gpt-6-1-sol has achieved a perfect score of 100.0% on the FrontierMath benchmark, establishing itself as the new leader in this category. This is a significant milestone for the model, which has also entered the top five on both the ARC-AGI-2 and GPQA Diamond benchmarks, showcasing its versatile capabilities across different evaluation metrics.
Additionally, claude-opus-5-5 and claude-sonnet-5-5 have both secured positions in the top five on the ARC-AGI-2 and GPQA Diamond benchmarks, respectively. The latter also made its mark by entering the top five on Humanity’s Last Exam, indicating a strong performance across multiple challenging tasks.
This round of evaluations also marks the first tracked measurements for several models, including claude-haiku-5-5, gpt-6-1-sol, claude-sonnet-5-5, gpt-6-luna, gpt-6-sol, claude-opus-5-5, and muse-spark-1-3. These initial measurements provide a baseline for future comparisons and highlight the growing diversity and capability of AI models in the industry.