Yandex AliceAI Cards - HumanEval: leaderboard
HumanEval Python function-completion problems, 5-shot, run by Yandex. Yandex AliceAI-T5-35B-A0.6B model card (September 2026); pretrained base checkpoints; score in percent (the card defines no metric), higher is better.
Metric: Score (%). Source: huggingface.co. Saturation forecast: Around February 2027. 6 models tracked.
Top models
| # | Model | Score | Overall rank |
|---|---|---|---|
| 1 | Qwen 3.5 35B A3B Base | 88.3 | #319 |
| 2 | Qwen3.5-4B-Base | 74.6 | #801 |
| 3 | Qwen3.5-2B-Base | 48.7 | #1288 |
| 4 | Gemma 4 E4B | 42.2 | #764 |
No result here: #3 Claude Opus 5.5, #5 GPT-6 Astra, #6 GPT-6.1 Sol.
Interactive version: theaggregate.ai/benchmark?slug=yandex-aliceai-cards-humaneval · How It Works · Data refreshed daily, snapshot 2026-11-07.