Hermes 2 Models by Nous Research
About
The Hermes 2 family of large language models, crafted by Nous Research, represents a suite of advanced models derived from diverse base architectures and refined for exceptional capabilities. Known for their superior performance across a wide spectrum of both general tasks and conversations, these models shine particularly in function calling and generating structured JSON outputs 23. Hermes 2 incorporates ChatML, facilitating structured multi-turn dialogues and specialized prompts tailored for dependable function calling and JSON output. Comprising models built on Llama 3 and Mistral architectures, they offer varying parameter configurations that cater to diverse performance needs 2. These models are evaluated highly in accuracy for function calling and JSON generation, making them apt for applications where reliable and interpretable machine-generated responses are crucial 4.
Current Variants
Use-when guidance is derived from seed capabilities, context, release, and replacement fields.
Use when the workload needs 8k context and 8B parameters.
Use when the workload needs 8k context and 8B parameters.
Use when the workload needs 32k context, 7B parameters, and function calling.
Use when the workload needs 32k context and 7B parameters.
Use when the workload needs 200k context and 34B parameters.
Use when the workload needs 4k context and 10.7B parameters.
Use when the workload needs 8k context and 70B parameters.
Use when the workload needs 8k context and 70B parameters.
Use when the workload needs 4k context and 70B parameters.
Use when the workload needs 4k context and 13B parameters.
Use when the workload needs 4k context and 7B parameters.
Use when the workload needs 33k context and 56B parameters.
Use when the workload needs 33k context and 56B parameters.
| Model | Use when | Released | Signals | Status |
|---|---|---|---|---|
| Hermes 2 Pro Llama 3 8B | Use when the workload needs 8k context and 8B parameters. | 2023-12 | 8k context8B parameters | Current |
| Hermes 2 Theta Llama 3 8B | Use when the workload needs 8k context and 8B parameters. | 2023-12 | 8k context8B parameters | Current |
| Hermes 2 Pro Mistral 7B | Use when the workload needs 32k context, 7B parameters, and function calling. | 2023-12 | 32k context7B parametersfunction calling | Current |
| Nous Hermes 2 Mixtral 8x7B | Use when the workload needs 32k context. | 2023-12 | 32k context | Current |
| Nous Hermes 2 Mistral 7B | Use when the workload needs 32k context and 7B parameters. | 2023-12 | 32k context7B parameters | Current |
| Nous Hermes 2 Yi 34B | Use when the workload needs 200k context and 34B parameters. | 2023-12 | 200k context34B parameters | Current |
| Nous Hermes 2 SOLAR 10.7B | Use when the workload needs 4k context and 10.7B parameters. | 2023-12 | 4k context10.7B parameters | Current |
| Hermes 2 Theta Llama 3 70B | Use when the workload needs 8k context and 70B parameters. | 2023-12 | 8k context70B parameters | Current |
| Hermes 2 Pro Llama 3 70B | Use when the workload needs 8k context and 70B parameters. | 2023-12 | 8k context70B parameters | Current |
| Nous Hermes 2 Llama 2 70B | Use when the workload needs 4k context and 70B parameters. | 2023-12 | 4k context70B parameters | Current |
| Nous Hermes 2 Llama 2 13B | Use when the workload needs 4k context and 13B parameters. | 2023-12 | 4k context13B parameters | Current |
| Nous Hermes 2 Llama 2 7B | Use when the workload needs 4k context and 7B parameters. | 2023-12 | 4k context7B parameters | Current |
| Together AI Nous-Hermes-2-Mixtral-8x7B-DPO | Use when the workload needs 33k context and 56B parameters. | 2023-12 | 33k context56B parameters | Current |
| OctoML Nous-Hermes-2-Mixtral-8x7B-DPO | Use when the workload needs 33k context and 56B parameters. | 2023-12 | 33k context56B parameters | Current |
Release Timeline
1 release groupSpecifications(14 models)
| Model | Released | Context | Parameters | Fn Calling |
|---|---|---|---|---|
| Hermes 2 Pro Llama 3 8B | 2023-12 | 8k | 8B | No |
| Hermes 2 Theta Llama 3 8B | 2023-12 | 8k | 8B | No |
| Hermes 2 Pro Mistral 7B | 2023-12 | 32k | 7B | Yes |
| Nous Hermes 2 Mixtral 8x7B | 2023-12 | 32k | 8x7B | No |
| Nous Hermes 2 Mistral 7B | 2023-12 | 32k | 7B | No |
| Nous Hermes 2 Yi 34B | 2023-12 | 200k | 34B | No |
| Nous Hermes 2 SOLAR 10.7B | 2023-12 | 4k | 10.7B | No |
| Hermes 2 Theta Llama 3 70B | 2023-12 | 8k | 70B | No |
| Hermes 2 Pro Llama 3 70B | 2023-12 | 8k | 70B | No |
| Nous Hermes 2 Llama 2 70B | 2023-12 | 4k | 70B | No |
| Nous Hermes 2 Llama 2 13B | 2023-12 | 4k | 13B | No |
| Nous Hermes 2 Llama 2 7B | 2023-12 | 4k | 7B | No |
| Together AI Nous-Hermes-2-Mixtral-8x7B-DPO | 2023-12 | 33k | 56B | No |
| OctoML Nous-Hermes-2-Mixtral-8x7B-DPO | 2023-12 | 33k | 56B | No |
Available From(9 providers)
Pricing
| Model | Provider | Input / 1M | Output / 1M | Type |
|---|---|---|---|---|
| Hermes 2 Theta Llama 3 8B | Replicate API | $0.05 | $0.25 | Serverless |
| Nous Hermes 2 SOLAR 10.7B | Replicate API | $0.1 | $0.5 | Serverless |
| Hermes 2 Pro Llama 3 8B | OpenRouter | $0.14 | $0.14 | Serverless |
| Hermes 2 Pro Llama 3 8B | Novita AI | $0.14 | $0.14 | Serverless |
| Nous Hermes 2 Mixtral 8x7B | OctoAI API (Deprecated) | $0.15 | $0.15 | Serverless |
| Hermes 2 Pro Llama 3 8B | OctoAI API (Deprecated) | $0.15 | $0.15 | Serverless |
| Hermes 2 Pro Mistral 7B | Fireworks AI | $0.2 | $0.2 | Provisioned |
| Nous Hermes 2 Mistral 7B | Together AI | $0.2 | $0.2 | Serverless |
| Nous Hermes 2 Yi 34B | Replicate API | $0.2 | $1 | Serverless |
| Hermes 2 Pro Llama 3 8B | Microsoft Foundry | $0.37 | $1.1 | Provisioned |
| Together AI Nous-Hermes-2-Mixtral-8x7B-DPO | Together AI | $0.4 | $0.4 | Serverless |
| OctoML Nous-Hermes-2-Mixtral-8x7B-DPO | OctoML (Deprecated) | $0.4 | $0.6 | Serverless |
| Nous Hermes 2 Mixtral 8x7B | Fireworks AI | $0.5 | $0.5 | Provisioned |
| Nous Hermes 2 Mixtral 8x7B | Together AI | $0.6 | $0.6 | Serverless |
| Nous Hermes 2 Yi 34B | Together AI | $0.8 | $0.8 | Serverless |
| Nous Hermes 2 Yi 34B | Fireworks AI | $0.9 | $0.9 | Provisioned |
Frequently Asked Questions
- What is Hermes 2 used for?
- Hermes 2 is used for agent workflows and tool use and structured outputs. The family description and listed model capabilities point to those workloads as the best fit.
- How does Hermes 2 compare to MOSS-Audio?
- Hermes 2 by Nous Research is strongest where you need agent workflows and tool use, while MOSS-Audio by MOSI Intelligence is the closest related family to check for multimodal. Hermes 2 has 14 listed variants and reaches up to 200k context, so compare the specs and pricing tables before choosing a production model.
- Which Hermes 2 model should I use?
- For the lowest listed input price, start with Hermes 2 Theta Llama 3 8B through Replicate API at $0.05/1M input tokens. For the most capable/latest local choice, evaluate Hermes 2 Pro Mistral 7B with 32k context and function calling.
Models(14)
Hermes 2 Pro Llama 3 8B
Hermes 2 Theta Llama 3 8B
Hermes 2 Pro Mistral 7B
Nous Hermes 2 Mixtral 8x7B
Nous Hermes 2 Mistral 7B
Nous Hermes 2 Yi 34B
Nous Hermes 2 SOLAR 10.7B
Hermes 2 Theta Llama 3 70B
Hermes 2 Pro Llama 3 70B
Nous Hermes 2 Llama 2 70B
Nous Hermes 2 Llama 2 13B
Nous Hermes 2 Llama 2 7B
Together AI Nous-Hermes-2-Mixtral-8x7B-DPO
OctoML Nous-Hermes-2-Mixtral-8x7B-DPO