Agent Memory Challenge 2026 Cycle 2 opens with $22K prize pool and 150-task coding track
omarsar0 · x · 2026-09-22
Agent Memory Challenge 2026 Cycle 2 is open, targeting the core problem of long-term memory: retrieving the right evidence and avoiding stale context rather than storing more history.
- Three tracks: Textual, Coding, Multimodal;
- Both open-source methods and commercial products compete via a shared Add/Search API with standardized evals and public results;
- Over $22,000 prize pool for eligible open-source teams.
Sergey Karayev recommends it to anyone working on agent memory: retrieval is the hardest memory problem since agents keep pulling stale context. The new cycle adds a coding track of 150 software tasks, each run with relevant history and again with noisy history.
More from coding & agent
- dotey's 4-step AI project workflow: PoC first, Claude Design, then lean MVP — dotey · 2026-09-22
- Musk shows Grok 4.7 doing real engineering at Tesla, shipping FSD features overnight — elonmusk · 2026-09-22
- Virtual Biotech teardown: 6 design decisions behind Science's 37,000-agent drug company — bravo_abad · 2026-09-22
- Half of 25 PM Job Openings Ask for Evals Experience; Ramp, Shopify, Harvey, Cursor Share ROI Data — lennysan · 2026-09-22
- 37,000-agent virtual drug company: the engineering behind Virtual Biotech — bravo_abad · 2026-09-22
- LinearSolveBench debuts: testing if AI can write fast C solvers for sparse linear systems — hgarud · 2026-09-22