← Back to People
Explained FrontierMath benchmark for measuring AI mathematical reasoning capabilities and discussed AI's advancement in solving PhD-level math problems.
4
Mentions
4
Articles
3
Outlets
Nov 2025 — Oct 2026
Coverage
Neutral
Sentiment
Associated Topics
Companies Covered
Coverage Patterns
How media typically covers Greg Burnham
Article Types
- news1
- analysis1
Media Angles
- technical2
Narrative Framing
- breakthrough1
- trend1
Appears In
Sentiment Distribution
Based on 4 scored articles
Neutral 25%Mixed 75%
Associated AI Models
GPT-5.21Claude Opus 4.61Gemini Deep Think1Aletheia1Claude1Gemini 31
Articles
Most recent firstArticles Written
Greg Burnham as author
Benchmark Scores = General Capability + Claudiness
Epoch AIanalysisneutralNov 24, 2025
Benchmark scores across AI models are dominated by a single 'General Capability' dimension, with a secondary 'Claudiness' dimension that captures Claude's unique performance profile across agentic tasks, vision, and math.
“Author of "Benchmark Scores = General Capability + Claudiness" in Epoch AI”
AI ResearchFoundation ModelsBenchmarkingGenerative AI
Also Mentioned In
Referenced in coverage
Can AI automate Epoch?
Epoch AIcautiousOct 9, 2026
“Greg Burnham co-authored the Epoch Automation Reports evaluating frontier and open-weight models on real work tasks.”
Research note: Have We Seen an Acceleration in Discoveries?
METRcautiousAug 24, 2026
“Provided helpful comments on METR acceleration research.”
How Far Can AI Go in Solving Math Mysteries?
IEEE SpectrumnewscautiousFeb 26, 2026
State-of-the-art AI models now solve over 40% of FrontierMath benchmark problems, up from 2% at launch, while Google DeepMind's Aletheia achieved publishable PhD-level mathematics autonomously, forcing the need for new, harder benchmarks.
“Explained FrontierMath benchmark for measuring AI mathematical reasoning capabilities and discussed AI's advancement in solving PhD-level math problems.”
AI ResearchAI SafetyFoundation ModelsBenchmarking and Evaluation