Frontier evals & benchmarks
MMLU/GPQA/SWE-bench/ARC-AGI/FrontierMath scoreboard
Benchmarks like SWE-bench, GPQA, MMLU, ARC-AGI-2, FrontierMath, and Humanity's Last Exam (with Epoch AI, LMArena, and Artificial Analysis as arbiters) are the shared scoreboard that drives lab competition and marketing. Benchmark saturation and contamination are now central credibility problems.
Key evals
SWE-bench, ARC-AGI-2, FrontierMath
Arbiters
Epoch, LMArena
How it fits the stack
Frontier evals & benchmarks with what it depends on (above) and what it feeds (below). The figure renders as a crawlable diagram and upgrades to an interactive 3D graph as it scrolls into view.
Frontier evals & benchmarks in the AI stack. Frontier evals & benchmarks with its immediate upstream dependencies (top) and downstream dependents (bottom) in the AI value chain. Hover a node in 3D, or read the full relationships below.
Graph data (text) — 2 entities, 1 relationships
- OpenAI —used by→ Frontier evals & benchmarks