GPT-OSS-20B
State‑of‑the‑art Mixture of Experts large language model with extended context length for text generation tasks.
GPT‑OSS‑20B is a 20.9B parameter Mixture of Experts (MoE) language model with 32 experts (4 active per token). It features an extended 131K context length with YARN rope scaling and uses a GPT‑4o compatible tokenizer. The model is quantized to MXFP4 for efficient on‑device deployment.
Not supported
This model is currently not supported on any IoT chipset.
To see performance metrics for this model on other chipsets, click the button below.
View for other chipsetsTechnical Details
Active experts per token:4
Model architecture:Mixture of Experts (MoE)
Number of experts:32
Number of parameters:20.91B
Response Rate:Rate of response generation after the first response token.
Supported languages:English
TTFT:Time To First Token is the time it takes to generate the first response token. This is expressed as a range because it varies based on the length of the prompt.
Tokenizer:BPE (GPT-2 style with GPT-4o preprocessing)
Applicable Scenarios
- Dialogue
- Content Generation
- Long Context Tasks
License
Model:APACHE-2.0
Terms of Use:Qualcomm® Generative AI usage and limitations
Tags
- llm
- generative-ai
- moe
Supported IoT Devices
- Dragonwing IQ-X7181
Supported IoT Chipsets
- Qualcomm® Dragonwing™ IQ-X7181
Looking for more? See models created by industry leaders.
Discover Model Makers