MemLens (64K) - Temporal Reasoning: leaderboard

Accuracy in percent on the MemLens 194 questions about dates, durations and the order of events, at a 64K-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 temporal reasoning (TR, 194 questions) at a 64K-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 May 2028. 27 models tracked. Updated 7 Oct 2026.

Top models

#ModelScoreOverall rank
1Qwen 3 VL 30B A3B Instruct55.15#903
2Qwen 3 VL 235B A22B Instruct54.12#541
3GLM-4.6V53.09#721
4Kimi K2.552.06#272
5Qwen 3.5 122B A10B51.55#350
6Qwen 3 VL 8B (Thinking)50.52#924
7Qwen 3 VL 8B Instruct47.42#1043
8Cosmos-Reason2-8B45.88#758
9Gemma 3 27B (IT)44.33#1341
10Gemma 3 12B (IT)44.33#1907
11Qwen 3 VL 235B A22B (Thinking)42.78#462
12Nemotron Nano 12B v2 VL42.78#2385
13Gemini 3.1 Pro (Preview)42.27#102
14GPT-5.442.27#138
15Qwen 3 VL 2B Instruct40.72#2389

No result here: #2 Claude Fable 5.2, #3 Gemini 4 Neon, #4 GPT-6.2 Astra.

Interactive version: theaggregate.ai/benchmark?slug=memlens-64k-temporal-reasoning · How It Works · Data refreshed daily, snapshot 2026-11-22.