MemLens (32K) - Knowledge Update: leaderboard
Accuracy in percent on the MemLens 116 questions whose answer changed in a later session, at a 32K-token interleaved image and text conversation history, with the evidence sessions hidden among distractor sessions (MemLens, 2026), judged by Qwen3-VL-235B. Higher is better.
Metric: Accuracy (%) on the knowledge update (KU, 116 questions) at a 32K-token history, interleaved multimodal multi-session conversation history with the evidence sessions hidden among haystack sessions, judged correct or incorrect by a Qwen3-VL-235B-A22B-Instruct judge; higher is better. Source: arxiv.org. Saturation forecast: Around September 2028. 27 models tracked. Updated 7 Oct 2026.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Kimi K2.5 | 50.86 | #266 |
| 2 | Gemini 3.1 Pro (Preview) | 49.14 | #99 |
| 3 | Qwen 3.5 122B A10B | 49.14 | #346 |
| 4 | GPT-5.4 | 47.41 | #135 |
| 5 | GLM-4.6V | 43.97 | #710 |
| 6 | Qwen 3.5 9B | 43.1 | #828 |
| 7 | Qwen 3.5 27B | 42.24 | #375 |
| 8 | Qwen 3 VL 235B A22B Instruct | 40.52 | #532 |
| 9 | Nemotron Nano 12B v2 VL | 39.66 | #2368 |
| 10 | Qwen 3 VL 30B A3B Instruct | 37.93 | #895 |
| 11 | Qwen 3 VL 235B A22B (Thinking) | 36.21 | #459 |
| 12 | Qwen 3.5 4B | 34.48 | #1241 |
| 13 | Qwen 3 VL 8B Instruct | 33.62 | #1029 |
| 14 | Qwen 3.5 2B | 33.62 | #5307 |
| 15 | Qwen 3 VL 4B (Thinking) | 33.62 | #1334 |
No result here: #1 Claude Fable 5.2, #3 Gemini 4 Neon, #4 GPT-6.2 Astra.
Interactive version: theaggregate.ai/benchmark?slug=memlens-32k-knowledge-update · How It Works · Data refreshed daily, snapshot 2026-11-21.