AnFeng_v3_Avocet: benchmark results
NLPark's AnFeng v3 Avocet, an early ~35B preview release in their v3 LLM series, SFT-tuned for Chinese and English. Provider: NLPark. Released 2024-04-30. Access: Open.
Aggregate Rating 1474 ± 23, rank #862 of 1635 rated models, from 19 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| Open LLM Leaderboard v1 - GSM8K | 74 | Accuracy (%) (5-shot) | 98.9 |
| Open LLM Leaderboard v1 - MMLU | 74.79 | Accuracy (%) (5-shot) | 96.3 |
| Open LLM Leaderboard v1 - HellaSwag | 87.43 | Normalized accuracy (%) (10-shot) | 86.1 |
| Open Chinese LLM - HellaSwag | 67.79 | Accuracy (%) | 84.9 |
| Open Chinese LLM - ARC Challenge | 59.3 | Accuracy (%) | 83.4 |
| Open LLM Leaderboard v1 - ARC Challenge | 67.75 | Normalized accuracy (%) (25-shot) | 78 |
| Open Chinese LLM Leaderboard | 63.67 | Official Nonzero-Task Mean (%) | 77.5 |
| Open LLM Leaderboard v1 - WinoGrande | 81.14 | Accuracy (%) (5-shot) | 77.4 |
| Open Chinese LLM - GSM8K | 55.88 | Accuracy (%) | 77.2 |
| Open LLM Leaderboard v1 - TruthfulQA MC2 | 58.63 | MC2 (%) (0-shot) | 72.4 |
| Open Chinese LLM - CMMLU | 61.84 | Accuracy (%) | 71.6 |
| Open Chinese LLM - TruthfulQA MC | 56.09 | Accuracy (%) | 65.7 |
Interactive version: theaggregate.ai/model?slug=anfeng-v3-avocet · How It Works · Data refreshed daily, snapshot 2026-11-05.