GET
/v1/schemas/together/chat/completions?kind=output
Accept:
| field | recorded source |
|---|---|
| schema | Source linked above. upstream-spec Source trace468edbdc879c48f500ca284270663ccc30cd83aee57a4e6a4e7a5bc04603b543{
"fetchedAt": 1786771359,
"extractorVersion": "1",
"verifiedAt": null
} |
Response body
7 properties · 7 required · 6 $defs| property | type | description | constraints |
|---|---|---|---|
| id* | string | ||
| choices* | object[] | ||
| prompt* | object[] | When `echo` is true, the prompt is included in the response. Additionally, when `logprobs` is also provided, log probability information is provided on the prompt. | |
| usage* | UsageData | ||
| created* | integer | ||
| model* | string | ||
| object* | const | The object type, which is always `text.completion`. | "text.completion" |
Types
6 $defsCompletionChoicesDataobject[]
| text | string | ||
| seed | integer | ||
| finish_reason | enum | stop | eos | length | tool_calls | function_call | |
| logprobs | LogprobsPart |
FinishReasonenum
stop | eos | length | tool_calls | function_call
LogprobsPartobject
| token_ids | number[] | List of token IDs corresponding to the logprobs | |
| tokens | string[] | List of token strings | |
| token_logprobs | number[] | List of token log probabilities | |
| top_logprobs | TopLogprobs | Top log probabilities for the tokens. |
PromptPartobject[]
| text | string | ||
| logprobs | LogprobsPart |
TopLogprobsobject
Top log probabilities for the tokens.
UsageDataobject
| prompt_tokens* | integer | ||
| completion_tokens* | integer | ||
| total_tokens* | integer |
Validate a payload
before you spend tokensshell
$ curl -X POST https://modelschemas.com/v1/validate \
-d '{"provider":"together","endpointId":"completions","payload":{…}}'
→ {"valid": false, "errors": [{"path": "#", "keyword": "required", …}]}