| Signal | Result | Reference |
|---|---|---|
| Tokenizer signature | Not measured | – |
| API surface | No other entry shares this exact contract | api_7b2b1f22 |
| Context and output | 16,385 context · 4,096 max output | declared |
| Reasoning contract | None declared | declared |
| Serving providers observed | None recorded yet | – |
This model offers four times the context length of gpt-3.5-turbo, allowing it to support approximately 20 pages of text in a single request at a higher cost. Training data: up...
Supported parameters
frequency_penaltylogit_biaslogprobsmax_completion_tokensmax_tokenspresence_penaltyresponse_formatseedstopstructured_outputstemperaturetool_choicetoolstop_logprobstop_p
defaults: {}
Cite this page as the evidence record for OpenAI: GPT-3.5 Turbo 16k: the URL is stable, measurements are dated, and revisions append to the history above. Method: tokenizer fingerprinting, v1, 50 strings.
