Fine Tuning

Fine TuningMethods

ModelsExpand Collapse

type DpoHyperparametersResp struct{…}

The hyperparameters used for the DPO fine-tuning job.

BatchSize DpoHyperparametersBatchSizeUnionRespoptional

Number of examples in each batch. A larger batch size means that model parameters are updated less frequently, but with lower variance.

Accepts one of the following:

type Auto string

int64

Beta DpoHyperparametersBetaUnionRespoptional

The beta value for the DPO method. A higher beta value will increase the weight of the penalty between the policy and reference model.

Accepts one of the following:

type Auto string

float64

LearningRateMultiplier DpoHyperparametersLearningRateMultiplierUnionRespoptional

Scaling factor for the learning rate. A smaller learning rate may be useful to avoid overfitting.

Accepts one of the following:

type Auto string

float64

NEpochs DpoHyperparametersNEpochsUnionRespoptional

The number of epochs to train the model for. An epoch refers to one full cycle through the training dataset.

Accepts one of the following:

type Auto string

int64

type DpoMethod struct{…}

Configuration for the DPO fine-tuning method.

Hyperparameters DpoHyperparametersRespoptional

The hyperparameters used for the DPO fine-tuning job.

type ReinforcementHyperparametersResp struct{…}

The hyperparameters used for the reinforcement fine-tuning job.

BatchSize ReinforcementHyperparametersBatchSizeUnionRespoptional

Number of examples in each batch. A larger batch size means that model parameters are updated less frequently, but with lower variance.

Accepts one of the following:

type Auto string

int64

ComputeMultiplier ReinforcementHyperparametersComputeMultiplierUnionRespoptional

Multiplier on amount of compute used for exploring search space during training.

Accepts one of the following:

type Auto string

float64

EvalInterval ReinforcementHyperparametersEvalIntervalUnionRespoptional

The number of training steps between evaluation runs.

Accepts one of the following:

type Auto string

int64

EvalSamples ReinforcementHyperparametersEvalSamplesUnionRespoptional

Number of evaluation samples to generate per training step.

Accepts one of the following:

type Auto string

int64

LearningRateMultiplier ReinforcementHyperparametersLearningRateMultiplierUnionRespoptional

Scaling factor for the learning rate. A smaller learning rate may be useful to avoid overfitting.

Accepts one of the following:

type Auto string

float64

NEpochs ReinforcementHyperparametersNEpochsUnionRespoptional

The number of epochs to train the model for. An epoch refers to one full cycle through the training dataset.

Accepts one of the following:

type Auto string

int64

ReasoningEffort ReinforcementHyperparametersReasoningEffortoptional

Level of reasoning effort.

Accepts one of the following:

const ReinforcementHyperparametersReasoningEffortDefault ReinforcementHyperparametersReasoningEffort = "default"

const ReinforcementHyperparametersReasoningEffortLow ReinforcementHyperparametersReasoningEffort = "low"

const ReinforcementHyperparametersReasoningEffortMedium ReinforcementHyperparametersReasoningEffort = "medium"

const ReinforcementHyperparametersReasoningEffortHigh ReinforcementHyperparametersReasoningEffort = "high"

type ReinforcementMethod struct{…}

Configuration for the reinforcement fine-tuning method.

Grader ReinforcementMethodGraderUnion

The grader used for the fine-tuning job.

Accepts one of the following:

type StringCheckGrader struct{…}

A StringCheckGrader object that performs a string comparison between input and reference using a specified operation.

Input string

The input text. This may include template strings.

Name string

The name of the grader.

Operation StringCheckGraderOperation

The string check operation to perform. One of eq, ne, like, or ilike.

Accepts one of the following:

const StringCheckGraderOperationEq StringCheckGraderOperation = "eq"

const StringCheckGraderOperationNe StringCheckGraderOperation = "ne"

const StringCheckGraderOperationLike StringCheckGraderOperation = "like"

const StringCheckGraderOperationIlike StringCheckGraderOperation = "ilike"

Reference string

The reference text. This may include template strings.

Type StringCheck

The object type, which is always string_check.

type TextSimilarityGrader struct{…}

A TextSimilarityGrader object which grades text based on similarity metrics.

EvaluationMetric TextSimilarityGraderEvaluationMetric

The evaluation metric to use. One of cosine, fuzzy_match, bleu, gleu, meteor, rouge_1, rouge_2, rouge_3, rouge_4, rouge_5, or rouge_l.

Accepts one of the following:

const TextSimilarityGraderEvaluationMetricCosine TextSimilarityGraderEvaluationMetric = "cosine"

const TextSimilarityGraderEvaluationMetricFuzzyMatch TextSimilarityGraderEvaluationMetric = "fuzzy_match"

const TextSimilarityGraderEvaluationMetricBleu TextSimilarityGraderEvaluationMetric = "bleu"

const TextSimilarityGraderEvaluationMetricGleu TextSimilarityGraderEvaluationMetric = "gleu"

const TextSimilarityGraderEvaluationMetricMeteor TextSimilarityGraderEvaluationMetric = "meteor"

const TextSimilarityGraderEvaluationMetricRouge1 TextSimilarityGraderEvaluationMetric = "rouge_1"

const TextSimilarityGraderEvaluationMetricRouge2 TextSimilarityGraderEvaluationMetric = "rouge_2"

const TextSimilarityGraderEvaluationMetricRouge3 TextSimilarityGraderEvaluationMetric = "rouge_3"

const TextSimilarityGraderEvaluationMetricRouge4 TextSimilarityGraderEvaluationMetric = "rouge_4"

const TextSimilarityGraderEvaluationMetricRouge5 TextSimilarityGraderEvaluationMetric = "rouge_5"

const TextSimilarityGraderEvaluationMetricRougeL TextSimilarityGraderEvaluationMetric = "rouge_l"

Input string

The text being graded.

Name string

The name of the grader.

Reference string

The text being graded against.

Type TextSimilarity

The type of grader.

type PythonGrader struct{…}

A PythonGrader object that runs a python script on the input.

Name string

The name of the grader.

Source string

The source code of the python script.

Type Python

The object type, which is always python.

ImageTag stringoptional

The image tag to use for the python script.

type ScoreModelGrader struct{…}

A ScoreModelGrader object that uses a model to assign a score to the input.

Input []ScoreModelGraderInput

The input messages evaluated by the grader. Supports text, output text, input image, and input audio content blocks, and may include template strings.

Content ScoreModelGraderInputContentUnion

Inputs to the model - can contain template strings. Supports text, output text, input images, and input audio, either as a single item or an array of items.

Accepts one of the following:

string

type ResponseInputText struct{…}

A text input to the model.

Text string

The text input to the model.

Type InputText

The type of the input item. Always input_text.

type ScoreModelGraderInputContentOutputText struct{…}

A text output from the model.

Text string

The text output from the model.

Type OutputText

The type of the output text. Always output_text.

type ScoreModelGraderInputContentInputImage struct{…}

An image input block used within EvalItem content arrays.

ImageURL string

The URL of the image input.

Type InputImage

The type of the image input. Always input_image.

Detail stringoptional

The detail level of the image to be sent to the model. One of high, low, or auto. Defaults to auto.

type ResponseInputAudio struct{…}

An audio input to the model.

InputAudio ResponseInputAudioInputAudio

Data string

Base64-encoded audio data.

Format string

The format of the audio data. Currently supported formats are mp3 and wav.

Accepts one of the following:

const ResponseInputAudioInputAudioFormatMP3 ResponseInputAudioInputAudioFormat = "mp3"

const ResponseInputAudioInputAudioFormatWAV ResponseInputAudioInputAudioFormat = "wav"

Type InputAudio

The type of the input item. Always input_audio.

type GraderInputs []GraderInputUnion

A list of inputs, each of which may be either an input text, output text, input image, or input audio object.

Accepts one of the following:

string

type ResponseInputText struct{…}

A text input to the model.

Text string

The text input to the model.

Type InputText

The type of the input item. Always input_text.

type GraderInputOutputText struct{…}

A text output from the model.

Text string

The text output from the model.

Type OutputText

The type of the output text. Always output_text.

type GraderInputInputImage struct{…}

An image input block used within EvalItem content arrays.

ImageURL string

The URL of the image input.

Type InputImage

The type of the image input. Always input_image.

Detail stringoptional

The detail level of the image to be sent to the model. One of high, low, or auto. Defaults to auto.

type ResponseInputAudio struct{…}

An audio input to the model.

InputAudio ResponseInputAudioInputAudio

Data string

Base64-encoded audio data.

Format string

The format of the audio data. Currently supported formats are mp3 and wav.

Accepts one of the following:

const ResponseInputAudioInputAudioFormatMP3 ResponseInputAudioInputAudioFormat = "mp3"

const ResponseInputAudioInputAudioFormatWAV ResponseInputAudioInputAudioFormat = "wav"

Type InputAudio

The type of the input item. Always input_audio.

Role string

The role of the message input. One of user, assistant, system, or developer.

Accepts one of the following:

const ScoreModelGraderInputRoleUser ScoreModelGraderInputRole = "user"

const ScoreModelGraderInputRoleAssistant ScoreModelGraderInputRole = "assistant"

const ScoreModelGraderInputRoleSystem ScoreModelGraderInputRole = "system"

const ScoreModelGraderInputRoleDeveloper ScoreModelGraderInputRole = "developer"

Type stringoptional

The type of the message input. Always message.

Model string

The model to use for the evaluation.

Name string

The name of the grader.

Type ScoreModel

The object type, which is always score_model.

Range []float64optional

The range of the score. Defaults to [0, 1].

SamplingParams ScoreModelGraderSamplingParamsoptional

The sampling parameters for the model.

MaxCompletionsTokens int64optional

The maximum number of tokens the grader model may generate in its response.

minimum1

ReasoningEffort ReasoningEffortoptional

Constrains effort on reasoning for reasoning models. Currently supported values are none, minimal, low, medium, high, and xhigh. Reducing reasoning effort can result in faster responses and fewer tokens used on reasoning in a response.

gpt-5.1 defaults to none, which does not perform reasoning. The supported reasoning values for gpt-5.1 are none, low, medium, and high. Tool calls are supported for all reasoning values in gpt-5.1.
All models before gpt-5.1 default to medium reasoning effort, and do not support none.
The gpt-5-pro model defaults to (and only supports) high reasoning effort.
xhigh is supported for all models after gpt-5.1-codex-max.

Seed int64optional

A seed value to initialize the randomness, during sampling.

Temperature float64optional

A higher temperature increases randomness in the outputs.

TopP float64optional

An alternative to temperature for nucleus sampling; 1.0 includes all tokens.

type MultiGrader struct{…}

A MultiGrader object combines the output of multiple graders to produce a single score.

CalculateOutput string

A formula to calculate the output based on grader results.

Graders MultiGraderGradersUnion