dolphin-2.9.3-Yi-1.5-34B-32k — benchmark results

Cognitive Computations' uncensored Dolphin 2.9.3 fine-tune of Yi 1.5 34B 32k with instruction, coding, agentic and function-calling data. Provider: Cognitive Computations. Released 2024-06-23. Access: Open.

Unified ELO 1472 ± 24, rank #900 of 1776 rated models, from 14 benchmark results.

Strongest benchmark results

BenchmarkScoreMetricPercentile
Open Chinese LLM - C-Eval Semantic89.16Accuracy (%)90.5
Open Chinese LLM - WinoGrande69.38Accuracy (%)90.1
Open Chinese LLM - CMMLU70.68Accuracy (%)87.8
Open Chinese LLM - GSM8K61.33Accuracy (%)86.1
Open LLM Leaderboard - GPQA12.42Score85.4
Open LLM Leaderboard - MMLU-Pro40.34Score85.2
Open Chinese LLM Leaderboard66.3Average Score (%)83.4
Open Chinese LLM - HellaSwag67.03Accuracy (%)82.8
Open LLM Leaderboard - BBH43.41Score82.6
Open Chinese LLM - ARC Challenge57.51Accuracy (%)75.2
Open LLM Leaderboard - MuSR13.35Score73.6
Open LLM Leaderboard - MATH Level 516.69Score65

Interactive version: theaggregate.ai/model?slug=dolphin-2-9-3-yi-1-5-34b-32k · How the rankings work · Data refreshed daily, snapshot 2026-07-22.