dolphin-2.9.1-yi-1.5-9B — benchmark results

Cognitive Computations' uncensored Dolphin fine-tune of Yi-1.5 9B for instruction, coding, and function-calling use. Provider: Cognitive Computations. Released 2024-05-18. Access: Open.

Unified ELO 1462 ± 16, rank #947 of 1776 rated models, from 15 benchmark results.

Strongest benchmark results

BenchmarkScoreMetricPercentile
Open Korean LLM Leaderboard469.53Average Score (%)84.3
Open Chinese LLM - WinoGrande67.56Accuracy (%)84.1
Open LLM Leaderboard - GPQA11.74Score83.7
Open Chinese LLM - ARC Challenge58.53Accuracy (%)80.7
Open Chinese LLM - GSM8K56.1Accuracy (%)77.4
Open Chinese LLM Leaderboard63.78Average Score (%)76.9
Open Chinese LLM - C-Eval Semantic85.03Accuracy (%)76.6
Open Chinese LLM - HellaSwag64.57Accuracy (%)76.6
Open Chinese LLM - CMMLU63.52Accuracy (%)74.8
Open LLM Leaderboard - MuSR13.52Score74.4
Open LLM Leaderboard - MMLU-Pro32.97Score73
Open LLM Leaderboard - BBH35.78Score72.6

Interactive version: theaggregate.ai/model?slug=dolphin-2-9-1-yi-1-5-9b · How the rankings work · Data refreshed daily, snapshot 2026-07-22.