Hermes 3 70B

Name: Hermes 3 70B
Author: Nous Research

Nous Research

Hermes 3 70B is the flagship instruction-following model by Nous Research, fine-tuned for advanced reasoning, creative writing, and complex task execution. It excels at instruction following and delivers high performance across a wide range of domains.

Key Specifications

Parameters

70.0B

Context

Release Date

August 14, 2024

Average Score

68.4%

Repository Model Weights Results Blog

Timeline

Key dates in the model's history

Announcement

August 14, 2024

Last Update

January 29, 2026

Today

October 5, 2026

Technical Specifications

Parameters

70.0B

Training Tokens

Knowledge Cutoff

Family

Fine-tuned from

llama-3.1-70b-instruct

Capabilities

MultimodalZeroEval

Benchmark Results

Model performance metrics across various tests and benchmarks

General Knowledge

Tests on general knowledge and understanding

HellaSwag

Accuracy • Self-reported

88.0%

Winogrande

Accuracy • Self-reported

83.0%

MMLU

Accuracy • Self-reported

79.1%

TruthfulQA

Self-reported by the model provider • Self-reported

63.3%

Mathematics

Mathematical problems and computations

MATH

Self-reported by the model provider • Self-reported

20.8%

Reasoning

Logical reasoning and analysis

GPQA

Self-reported by the model provider • Self-reported

66.1%

Other Tests

Specialized benchmarks

MT-Bench

Score 100 • Self-reported

89.9%

BoolQ

Accuracy • Self-reported

88.0%

PIQA

Accuracy • Self-reported

84.0%

ARC-E

Accuracy • Self-reported

83.0%

AGIEval

Self-reported by the model provider • Self-reported

56.2%

ARC-C

Self-reported by the model provider • Self-reported

65.5%

BBH

Self-reported by the model provider • Self-reported

67.8%

IFBench

Self-reported by the model provider • Self-reported

81.2%

MMLU-Pro

Self-reported by the model provider • Self-reported

47.2%

MuSR

Self-reported by the model provider • Self-reported

50.7%

OpenBookQA

Self-reported by the model provider • Self-reported

49.4%

License & Metadata

License

apache-2.0

Announcement Date

August 14, 2024

Last Updated

January 29, 2026

Similar Models

All Models

GLM-4.7-Flash

Zhipu AI

30.0B

Best score:0.8 (TAU)

Released:Jan 2026

Price:$0.07/1M tokens

Llama 3.1 70B Instruct

Gemma 2 27B

Google

27.2B

Best score:0.8 (MMLU)

Released:Jun 2024

Llama 3.1 Nemotron 70B Instruct

NVIDIA

70.0B

Best score:0.8 (MMLU)

Released:Oct 2024

Phi 4

Microsoft

14.7B

Best score:0.8 (MMLU)

Released:Dec 2024

Price:$0.07/1M tokens

Phi-3.5-MoE-instruct

Microsoft

60.0B

Best score:0.9 (ARC)

Released:Aug 2024

Magistral Small 2506

Mistral AI

24.0B

Best score:0.7 (GPQA)

Released:Jun 2025

IBM Granite 4.2 30B

IBM

30.0B

Best score:0.7 (GPQA)

Released:Aug 2026

Recommendations are based on similarity of characteristics: developer organization, multimodality, parameter size, and benchmark performance. Choose a model to compare or go to the full catalog to browse all available AI models.