Yandex AliceAI Cards - LongMemEval 128K: leaderboard
LongMemEval questions that need information from long dialogue histories, at 128K tokens, 5-shot, run by Yandex. Yandex AliceAI-Foundation-80B-A3B-Base model card (September 2026); pretrained base checkpoints, greedy decoding; accuracy in percent, higher is better.
Metric: Accuracy (%). Source: huggingface.co. Saturation forecast: Around February 2027. 5 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Qwen 3.5 35B A3B Base | 55.6 | #705 |
No result here: #4 Claude Opus 5.5, #5 GPT-6 Astra, #9 GPT-6.1 Sol.
Interactive version: theaggregate.ai/benchmark?slug=yandex-aliceai-cards-longmemeval-128k · How It Works · Data refreshed daily, snapshot 2026-11-08.