EpochAI's Tier 4 benchmark goes from 5% to 98% top score in under 14 months, now saturated
teortaxesTex · x · 2026-09-11
EpochAI reports that when its Tier 4 benchmark launched on July 11, 2025, the top score was just 5%. Less than 14 months later, the top score stands at 98%, and the organization now considers the benchmark saturated — a stark illustration of how quickly frontier models exhaust difficult evals and a fresh challenge for benchmark design.
Related event: Epoch AI's Tier 4 Math Benchmark Saturated in 14 Months(3 posts)→
More from Models
- OpenAI reportedly has a next-gen model significantly more capable than GPT-6 Astra — sjgadler · 2026-09-11
- FrontierMath Tier 4, once the hardest math benchmark, falls after 1.5 years — Jsevillamol · 2026-09-11
- Octen Search debuts third on search index at 16.9s per task, $0.058 cost — ArtificialAnlys · 2026-09-11
- ValsAI: Astra first AI to reach Minecraft Nether Fortress in long-horizon eval — scaling01 · 2026-09-11
- Dev: You Can Tell Who Has a Real RL Pipeline Just From Model Outputs — teortaxesTex · 2026-09-11
- DeepSeek Flash impresses: non-sycophantic, argumentative, and blazing fast — oran_ge · 2026-09-11