AlpsBench - Retrieval (100 Distractors): leaderboard
Metric: Recall (0-1, times 100): share of queries for which the model selects the human-labelled relevant memory from one positive and 100 randomly sampled distractor memories; AlpsBench's 2,500 human-verified instances per task built from long-term real user-LLM dialogues in WildChat; higher is better. Source: arxiv.org. Saturation forecast: Estimated already saturated. 7 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | DeepSeek V3.2 (Thinking) | 97.28 | #198 (DeepSeek V3.2) |
| 2 | Claude Sonnet 4.5 | 96.91 | #138 |
| 3 | Gemini 3 Flash (Preview) | 96.42 | #78 |
| 4 | GPT-5.2 | 95.19 | #105 |
| 5 | Qwen 3 Max | 94.02 | #201 |
| 6 | GPT-4.1 Mini | 91.1 | #346 |
| 7 | Llama 4 Maverick | 89.58 | #451 |
No result here: #3 Claude Opus 5.5, #5 GPT-6 Astra, #8 Claude Fable 5.1.
Interactive version: theaggregate.ai/benchmark?slug=alpsbench-retrieval-100-distractors · How It Works · Data refreshed daily, snapshot 2026-10-11.