LLM built for 12M-token reasoning

(subq.ai) by tschibrillo | May 10, 2026 | 0 comments on HN