Fetching from the wire…
01
02
03
04
05
06
07
08
Public story · 2026-07-23 · high
It hits up to 76.1% pass@1 on four to six times fewer tokens than summarization, with the code public on GitHub.
Why now: The paper landed the same week a senior Anthropic engineer's graph-engineering thread went viral and Andrew Ng's team shipped a course on the same unbenchmarked idea.
PRO-LONG skips the summarization step every agent-memory system leans on, storing the full log instead and letting the agent grep it, per the paper.
That matters for anyone building agent memory. PRO-LONG beats a base coding agent by 18.0 percentage points across frontier models on the full ARC-AGI-3 public game set. It hits up to 76.1% pass@1 while spending 4.2 to 5.8 times fewer tokens than a summarize-and-evict setup. Run with Fable 5, it reaches 97.4% best@2 for $1,750 total. The code is public on GitHub.
The logic holds up once you sit with it. Summarization is lossy compression done before you know what you'll need. You're guessing at write time what'll matter at query time. Grep is lossless storage with retrieval deferred until the agent actually needs the answer. That's also where the token savings come from. You're not paying to re-read a summary every turn. You're paying only for the three lines you pulled out.
This landed the same week a senior Anthropic engineer's 12-page treatment of graph engineering went viral, four threads in 48 hours. It argues agent state needs an explicit typed node graph instead of a log. Andrew Ng's team shipped a course on the same idea days later. Turing Post is already running a piece asking if the whole thing is real. A definitive-guide post and a skeptic piece landing in the same stretch is what a hype phase looks like, not a settled one. PRO-LONG skipped that step and shipped a benchmark instead. A related paper, ContextSniper, points the same direction. Filtering verbose tool output cut Claude Code's tokens 38.9% on SWE-bench Lite tasks with resolution rates essentially unchanged.
Each link below shares sources, entities, or timing with this story.
Claude Opus built by Anthropic / Shared entities / Same source domain / Shared topic / Earlier coverage
Linked by a graph relationship (Claude Opus built by Anthropic); both cover Anthropic, SWE, When; reported by the same outlet (arxiv.org).
Anthropic partners with Microsoft / Shared entities / Same source domain / Shared topic / Earlier coverage
Linked by a graph relationship (Anthropic partners with Microsoft); both cover AGI, ARC, Code, SWE; reported by the same outlet (arxiv.org).
Anthropic released Claude Code / Shared entities / Shared topic / Earlier coverage
Linked by a graph relationship (Anthropic released Claude Code); both cover Anthropic, Claude Code, OpenClaw, When; overlapping topics (agent, code, point).
Linked by a graph relationship (Anthropic released Claude Code); both cover Claude Code, Most, Treat; overlapping topics (agent, context, engineering, point, token).
Anthropic released Claude Code / Shared entities / Same source domain / Shared topic / Earlier coverage
Linked by a graph relationship (Anthropic released Claude Code); both cover Claude Code, Keep, Points; reported by the same outlet (arxiv.org).
Anthropic released Claude Code / Shared entities / Shared topic / Earlier coverage
Linked by a graph relationship (Anthropic released Claude Code); both cover Claude Code, Code, Most, When; overlapping topics (agent, code).
Linked by a graph relationship (Anthropic released Claude Code); both cover Anthropic, Claude Code, Code, When; overlapping topics (agent, code).
Anthropic released Claude / Shared entities / Same source domain / Shared topic / Earlier coverage / Tension
Linked by a graph relationship (Anthropic released Claude); both cover Anthropic, Keep; reported by the same outlet (arxiv.org).