Skip to main content
Summarization models take long passages and return a short, faithful summary. Call either the Responses API (POST /v1/responses) or the OpenAI-compatible Chat Completions API (POST /v1/chat/completions).

llama-3.1-8b-instruct-fast

Meta’s Llama 3.1 Instruct, tuned for fast, low-cost summarization at scale on the ZeroGPU edge network. Its 128K-token context window takes in entire documents, long transcripts, and full email or chat threads in a single pass. No chunking, no stitching partial summaries back together. Built for the high-volume summarization that runs around your core product: condensing long reports and transcripts, summarizing email and chat threads, and processing articles in bulk. When context length matters and you’re summarizing on every request, this is the model that keeps it fast and affordable.
References: Model docsTermsPrivacy
The summary is returned as output[].content[].text (Responses API) or choices[].message.content (Chat Completions).
Response

t5-small

Google’s T5-small introduced the text-to-text paradigm that changed how the industry thinks about NLP. At 60M parameters, it’s the most versatile encoder-decoder model in its class, pre-trained on the Colossal Clean Crawled Corpus (C4) and fine-tunable for virtually any text task with minimal data. Its encoder-decoder architecture makes it especially strong at tasks that require both understanding and generation, like summarization and rewriting. A proven workhorse that’s been battle-tested across millions of production deployments.
References: Model cardTermsPrivacy Inputs are truncated at 512 tokens, so chunk anything longer, or use llama-3.1-8b-instruct-fast when the whole document has to land in one pass. Summaries are extractive in style: T5 compresses the source sentences rather than rewriting them.
Response