Letta Memory Benchmark: leaderboard
Each model runs as a Letta agent with memory tools on four tasks: archival memory read, core memory read, core memory update and a hard core memory write, each over 100 items repeated 3 times; the score is the published mean of the four task scores. Unlike Letta's Context-Bench, which grades file and skill use, it tests the agent's own memory blocks.
Metric: Average score (%) over four Letta memory tasks, 100 items each, 3 repeats, Letta agent framework. Source: raw.githubusercontent.com. Saturation forecast: Estimated already saturated. 18 models tracked. Updated 16 Nov 2026.
Top models
| # | Model | Setting | Score | Overall rank |
|---|---|---|---|---|
| 1 | Claude Sonnet 4 | Thinking | 92.08 | #406 |
| 2 | Claude Sonnet 4 (20250514) | 91.25 | #388 | |
| 3 | GPT-4o (2024-08-06) | 90.5 | #736 | |
| 4 | GPT-4o Mini (2024-07-18) | 87.83 | #1511 | |
| 5 | Gemini 2.5 Flash (Preview 05-20) | 87.58 | #423 | |
| 6 | Gemini 2.5 Pro (Preview 05-06) | 84.92 | #279 | |
| 7 | GPT-4.1 Mini | 80.17 | #781 | |
| 8 | Claude 3.7 Sonnet (20250219) | 78.58 | #414 | |
| 9 | Claude 3.5 Haiku (20241022) | 70.5 | #1601 | |
| 10 | Gemini 2.5 Flash (Preview 04-17) | 67.92 | #458 | |
| 11 | O4 Mini | 66.5 | #356 | |
| 12 | Llama 4 Maverick Instruct FP8 | 61.25 | #1010 | |
| 13 | Llama 3.3 70B Instruct | 55.67 | #1379 | |
| 14 | GPT-4.1 Nano | 47.75 | #2035 | |
| 15 | Llama 4 Scout Instruct | 47 | #1579 |
No result here: #1 Claude Fable 5.2, #2 Gemini 4 Carbon, #5 Claude Opus 5.5.
Interactive version: theaggregate.ai/benchmark?slug=letta-memory-benchmark · How It Works · Data refreshed daily, snapshot 2026-11-16.