MemLens (32K) - Information Extraction: leaderboard
Accuracy in percent on the MemLens 246 questions that extract a fact or image detail from one earlier 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 information extraction (IE, 246 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 July 2027. 26 models tracked. Updated 7 Oct 2026.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Qwen 3.5 122B A10B | 74.39 | #346 |
| 2 | Qwen 3.5 27B | 70.33 | #375 |
| 3 | GPT-5.4 | 69.51 | #135 |
| 4 | Qwen 3.5 9B | 67.48 | #828 |
| 5 | Qwen 3 VL 30B A3B Instruct | 65.45 | #895 |
| 6 | Qwen 3.5 2B | 64.23 | #5307 |
| 7 | Nemotron Nano 12B v2 VL | 63.82 | #2368 |
| 8 | GLM-4.5V | 61.79 | #754 |
| 9 | Qwen 3 VL 235B A22B Instruct | 60.98 | #532 |
| 10 | GLM-4.6V | 59.35 | #710 |
| 11 | Gemini 3.1 Pro (Preview) | 57.32 | #99 |
| 12 | Qwen 3.5 4B | 56.91 | #1241 |
| 13 | Cosmos-Reason2-8B | 55.69 | #755 |
| 14 | Qwen 3 VL 8B Instruct | 52.85 | #1029 |
| 15 | Qwen 3 VL 4B Instruct | 52.03 | #1332 |
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-information-extraction · How It Works · Data refreshed daily, snapshot 2026-11-21.