Anthropic posts Claude-discovered subexponential 3-SAT algorithm refuting ETH on Hexagon
thebasepoint · x · 2026-10-27
On the new Hexagon platform for AI-generated math results, Anthropic contributors published a deterministic 2^{o(n)}-time algorithm for 3-SAT that in fact solves the harder #3-SAT (counting satisfying assignments), refuting the Exponential Time Hypothesis.
- The core algorithm was discovered Oct 16 by an internal research Claude model with 14M output tokens and no human input; formalized in Lean 4.
- The paper is flagged "primarily AI-generated" with human understanding marked as partial; Anthropic researchers polished the presentation.
- The authors note the result doesn't bear on P vs NP directly and yields no practical fast algorithms.
- The provenance statement notes OpenAI's internal model independently and concurrently claims a similar result.
- The posted introduction recaps prior records: O(1.328^n) deterministic and O(1.308^n) randomized 3-SAT algorithms.
More from Models
- OpenAI claims models refuted the Exponential Time Hypothesis, proved Kakeya, with Lean proofs — LinusEkenstam · 2026-10-27
- Rumor: OpenAI's 25-page proof of Sogge smoothing conjecture bypasses 20 years of work — teortaxesTex · 2026-10-27
- Cohere releases Qwen traces: anomaly is real, but backdoor intent claims have no evidence — basedjensen · 2026-10-27
- Fields medalist Yu Deng calls out OpenAI's double standard on distillation — teortaxesTex · 2026-10-27
- Commentary: OpenAI won't release reasoning traces because the reasoning process is the real product — max_paperclips · 2026-10-27
- 229k-sample replication of Cohere's Qwen 3.8 malicious PUT claim raises more questions — builderjaydub · 2026-10-27