Stateless AI agents cut long-horizon research tokens
Why it matters
SLA keeps research state in the harness and gives each agent a fresh context. On Anthropic kernel optimization with Codex, it matched the strongest baseline with 93.1% fewer tokens (67.9M vs 986.3M). Best full-budget result on every task; Advisor used 0.24–0.51% of tokens.
Read the source: arXiv (opens in a new tab)
All hits from 13 Oct 2026More AgentsGet the daily hits by email