airoboros Models by Jon Durbin
About
The Airoboros family of large language models, developed by Jon Durbin, is designed to enhance instruction-following capabilities across diverse tasks such as question answering, summarization, and code generation. These models are fine-tuned using a self-instructing method and built upon various base models, including Llama 2 and MPT-30B, with sizes ranging from 7B to 70B parameters. Quantized versions optimized for different hardware setups are available on Hugging Face. The Airoboros project also includes tools for creating customized datasets, enabling the development of specialized expert models. The licensing is intricate, involving a custom license alongside the Meta Llama 2 license.
Current Variants
Use-when guidance is derived from seed capabilities, context, release, and replacement fields.
Use when the workload needs 70B parameters and structured outputs.
| Model | Use when | Released | Signals | Status |
|---|---|---|---|---|
| airoboros L2 70B 2.2.1 | Use when the workload needs 70B parameters and structured outputs. | 2023-07 | 70B parametersstructured outputs | Current |
| airoboros L2 13B 2.2.1 | Use when the workload needs 13B parameters. | 2023-07 | 13B parameters | Current |
| airoboros L2 7B 2.2.1 | Use when the workload needs 7B parameters. | 2023-07 | 7B parameters | Current |
Release Timeline
1 release groupSpecifications(3 models)
| Model | Released | Parameters | Structured Outputs |
|---|---|---|---|
| airoboros L2 70B 2.2.1 | 2023-07 | 70B | Yes |
| airoboros L2 13B 2.2.1 | 2023-07 | 13B | No |
| airoboros L2 7B 2.2.1 | 2023-07 | 7B | No |
Available From(1 provider)
Pricing
| Model | Provider | Input / 1M | Output / 1M | Type |
|---|---|---|---|---|
| airoboros L2 70B 2.2.1 | DeepInfra | $0.45 | $0.65 | Serverless |
Frequently Asked Questions
- What is airoboros used for?
- airoboros is used for structured outputs and coding. The family description and listed model capabilities point to those workloads as the best fit.
- How does airoboros compare to Claude 3?
- airoboros by Jon Durbin is strongest where you need structured outputs, while Claude 3 by Anthropic is the closest related family to check for vision and multimodal work. airoboros has 3 listed variants, while Claude 3 reaches up to 200k context, so compare the specs and pricing tables before choosing a production model.
- Which airoboros model should I use?
- For the lowest listed input price, start with airoboros L2 70B 2.2.1 through DeepInfra at $0.45/1M input tokens. For the most capable/latest local choice, evaluate airoboros L2 70B 2.2.1 with structured outputs.