Meta Muse Wonder Scorecard - MemoryBench: leaderboard
Meta's internal MemoryBench: recalling critical details across long multi-turn conversations of 64K to 100K tokens, with distractor conversations and turns that conflict with or update earlier facts; an LLM judge scores each answer against weighted, question-specific rubrics. The questions are not public. Meta ran this for its October 12, 2026 Muse Wonder release scorecard, with every model at max reasoning effort; competitor runs are Meta's best-effort internal runs. Metric: Rubric score (%). Higher is better. Source: https://research.meta.ai/blog/muse-wonder
Metric: Rubric score (%). Source: research.meta.ai. Saturation forecast: Around April 2027. 5 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Muse Wonder (Max) | 85.6 | #14 (Muse Wonder) |
| 2 | Muse Wonder Full (Max) | 85.5 | #10 (Muse Wonder Full) |
| 3 | Claude Opus 5.5 (Max) | 81.5 | #4 (Claude Opus 5.5) |
| 4 | GPT-6 Astra (Max) | 81.4 | #6 (GPT-6 Astra) |
| 5 | Claude Fable 5.1 (Max) | 79 | #9 (Claude Fable 5.1) |
No result here: #7 GPT-6.1 Sol, #8 Gemini 4 Argon.
Interactive version: theaggregate.ai/benchmark?slug=meta-muse-wonder-scorecard-memorybench · How It Works · Data refreshed daily, snapshot 2026-10-18.