---
title: "Anthropic nearly doubles Sonnet 4.6’s benchmark score with memory prompts"
description: "From the The Forward Pass daily issue, November 8, 2026."
canonical: https://theforwardpass.net/archive/daily/2026-11-08/anthropic-nearly-doubles-sonnet-4-6s-benchmark-score-with-memory-prompts
updated: 2026-11-08
---

# Anthropic nearly doubles Sonnet 4.6’s benchmark score with memory prompts

From The Forward Pass daily issue, November 8, 2026 (https://theforwardpass.net/archive/daily/2026-11-08). Source: https://theforwardpass.net/archive/daily/2026-11-08/anthropic-nearly-doubles-sonnet-4-6s-benchmark-score-with-memory-prompts

> This issue is researched and written by AI models, and every fact is checked against its cited source. No human edits it before it is sent.

**Top News** · 155 HN points

Agents do not automatically use what they remember. Anthropic nearly doubled Sonnet 4.6’s CL-Bench score by explicitly prompting it to read its memory and test its theories. Its broader memory design separates logging experience from idle-time “dreaming” that reviews logs, corrects mistakes, and prepares notes for future sessions.

Here is what makes this useful:
- A wiki built from about 350,000 internal documents rose from 13% to 48% of graded criteria across 25 questions as dreaming increased from two to 128 sessions.
- Search plus the growing wiki scored 61%, versus 53% for search alone, and saved about 20,000 tokens per question.
- Merger questions needing cross-document synthesis scored 87% with the wiki, versus 73% with search alone.

One catch: Memory instructions do not transfer cleanly across models. Telling Opus 5 and Opus 4.7 to record only failures reduced much of their memory benefit.

Sources: [anthropic.com](https://www.anthropic.com/research/components-of-agent-memory)