Yandex AliceAI Cards - IMO-AnswerBench (pass@8): leaderboard
IMO-AnswerBench olympiad problems with short verifiable answers, 0-shot; pass@8 counts a problem solved when any of 8 samples is correct. Yandex AliceAI-Foundation-80B-A3B-Base model card (September 2026), complex-reasoning table; pretrained base checkpoints sampled at temperature 1; percent, higher is better.
Metric: Pass@8 (%). Source: huggingface.co. Saturation forecast: Around February 2027. 3 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Qwen 3.5 35B A3B Base | 84.5 | #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-imo-answerbench-pass-8 · How It Works · Data refreshed daily, snapshot 2026-11-08.