AxModelInfo

type AxModelInfo = object;

Defined in: https://github.com/ax-llm/ax/blob/3d5aa8dac28a2637e4150f022dcd02e6fc3309f0/src/ax/ai/types.ts#L70

Properties

aliases?

optional aliases: string[];

Defined in: https://github.com/ax-llm/ax/blob/3d5aa8dac28a2637e4150f022dcd02e6fc3309f0/src/ax/ai/types.ts#L99

audio?

optional audio: object;

Defined in: https://github.com/ax-llm/ax/blob/3d5aa8dac28a2637e4150f022dcd02e6fc3309f0/src/ax/ai/types.ts#L146

input?

optional input: boolean;

output?

optional output: boolean;

cacheReadTokenCostPer1M?

optional cacheReadTokenCostPer1M: number;

Defined in: https://github.com/ax-llm/ax/blob/3d5aa8dac28a2637e4150f022dcd02e6fc3309f0/src/ax/ai/types.ts#L76

cacheWriteTokenCostPer1M?

optional cacheWriteTokenCostPer1M: number;

Defined in: https://github.com/ax-llm/ax/blob/3d5aa8dac28a2637e4150f022dcd02e6fc3309f0/src/ax/ai/types.ts#L77

characterIsToken?

optional characterIsToken: boolean;

Defined in: https://github.com/ax-llm/ax/blob/3d5aa8dac28a2637e4150f022dcd02e6fc3309f0/src/ax/ai/types.ts#L73

completionTokenCostPer1M?

optional completionTokenCostPer1M: number;

Defined in: https://github.com/ax-llm/ax/blob/3d5aa8dac28a2637e4150f022dcd02e6fc3309f0/src/ax/ai/types.ts#L75

contextWindow?

optional contextWindow: number;

Defined in: https://github.com/ax-llm/ax/blob/3d5aa8dac28a2637e4150f022dcd02e6fc3309f0/src/ax/ai/types.ts#L157

currency?

optional currency: string;

Defined in: https://github.com/ax-llm/ax/blob/3d5aa8dac28a2637e4150f022dcd02e6fc3309f0/src/ax/ai/types.ts#L72

dataUse?

optional dataUse: object;

Defined in: https://github.com/ax-llm/ax/blob/3d5aa8dac28a2637e4150f022dcd02e6fc3309f0/src/ax/ai/types.ts#L162

Provider data-use policy that materially differs between model variants.

appliesTo?

optional appliesTo: readonly ("prompt" | "completion")[];

providerTraining

providerTraining: "not-used" | "allowed";

deprecatedOn?

optional deprecatedOn: string;

Defined in: https://github.com/ax-llm/ax/blob/3d5aa8dac28a2637e4150f022dcd02e6fc3309f0/src/ax/ai/types.ts#L160

ISO date (YYYY-MM-DD) the upstream provider will stop serving this model.

fastCacheReadTokenCostPer1M?

optional fastCacheReadTokenCostPer1M: number;

Defined in: https://github.com/ax-llm/ax/blob/3d5aa8dac28a2637e4150f022dcd02e6fc3309f0/src/ax/ai/types.ts#L91

Cache read token cost when a provider speed tier such as Anthropic fast mode is active.

fastCacheWriteTokenCostPer1M?

optional fastCacheWriteTokenCostPer1M: number;

Defined in: https://github.com/ax-llm/ax/blob/3d5aa8dac28a2637e4150f022dcd02e6fc3309f0/src/ax/ai/types.ts#L93

Cache write token cost when a provider speed tier such as Anthropic fast mode is active.

fastCompletionTokenCostPer1M?

optional fastCompletionTokenCostPer1M: number;

Defined in: https://github.com/ax-llm/ax/blob/3d5aa8dac28a2637e4150f022dcd02e6fc3309f0/src/ax/ai/types.ts#L89

Completion token cost when a provider speed tier such as Anthropic fast mode is active.

fastPromptTokenCostPer1M?

optional fastPromptTokenCostPer1M: number;

Defined in: https://github.com/ax-llm/ax/blob/3d5aa8dac28a2637e4150f022dcd02e6fc3309f0/src/ax/ai/types.ts#L87

Prompt token cost when a provider speed tier such as Anthropic fast mode is active.

isDeprecated?

optional isDeprecated: boolean;

Defined in: https://github.com/ax-llm/ax/blob/3d5aa8dac28a2637e4150f022dcd02e6fc3309f0/src/ax/ai/types.ts#L158

isExpensive?

optional isExpensive: boolean;

Defined in: https://github.com/ax-llm/ax/blob/3d5aa8dac28a2637e4150f022dcd02e6fc3309f0/src/ax/ai/types.ts#L156

longContextCacheReadTokenCostPer1M?

optional longContextCacheReadTokenCostPer1M: number;

Defined in: https://github.com/ax-llm/ax/blob/3d5aa8dac28a2637e4150f022dcd02e6fc3309f0/src/ax/ai/types.ts#L83

Cache read token cost for requests exceeding longContextThreshold

longContextCacheWriteTokenCostPer1M?

optional longContextCacheWriteTokenCostPer1M: number;

Defined in: https://github.com/ax-llm/ax/blob/3d5aa8dac28a2637e4150f022dcd02e6fc3309f0/src/ax/ai/types.ts#L98

longContextCompletionTokenCostPer1M?

optional longContextCompletionTokenCostPer1M: number;

Defined in: https://github.com/ax-llm/ax/blob/3d5aa8dac28a2637e4150f022dcd02e6fc3309f0/src/ax/ai/types.ts#L81

Completion token cost for requests exceeding longContextThreshold

longContextPromptTokenCostPer1M?

optional longContextPromptTokenCostPer1M: number;

Defined in: https://github.com/ax-llm/ax/blob/3d5aa8dac28a2637e4150f022dcd02e6fc3309f0/src/ax/ai/types.ts#L79

Prompt token cost for requests exceeding longContextThreshold

longContextThreshold?

optional longContextThreshold: number;

Defined in: https://github.com/ax-llm/ax/blob/3d5aa8dac28a2637e4150f022dcd02e6fc3309f0/src/ax/ai/types.ts#L85

Total input token count (including cached) above which long-context rates apply

maxTokens?

optional maxTokens: number;

Defined in: https://github.com/ax-llm/ax/blob/3d5aa8dac28a2637e4150f022dcd02e6fc3309f0/src/ax/ai/types.ts#L155

modalities?

optional modalities: object;

Defined in: https://github.com/ax-llm/ax/blob/3d5aa8dac28a2637e4150f022dcd02e6fc3309f0/src/ax/ai/types.ts#L151

Exact provider-declared model input and output modalities.

input

input: readonly ("text" | "image" | "audio" | "video" | "pdf")[];

output

output: readonly ("text" | "image")[];

name

name: string;

Defined in: https://github.com/ax-llm/ax/blob/3d5aa8dac28a2637e4150f022dcd02e6fc3309f0/src/ax/ai/types.ts#L71

notSupported?

optional notSupported: object;

Defined in: https://github.com/ax-llm/ax/blob/3d5aa8dac28a2637e4150f022dcd02e6fc3309f0/src/ax/ai/types.ts#L132

Sampling parameters the model rejects. With supported.samplingWithoutReasoning it rejects them only while it reasons. Ax never sends a default value for them; it sends an explicit value when the model accepts it for that request and otherwise drops it with a one-time warning.

frequencyPenalty?

optional frequencyPenalty: boolean;

presencePenalty?

optional presencePenalty: boolean;

temperature?

optional temperature: boolean;

temperatureWithTopP?

optional temperatureWithTopP: boolean;

The model takes temperature or topP but not both in one request. An explicit topP with only the default temperature is sent alone; with an explicit temperature too, Ax keeps the temperature and drops topP with a one-time warning.

topK?

optional topK: boolean;

topP?

optional topP: boolean;

promptTokenCostPer1M?

optional promptTokenCostPer1M: number;

Defined in: https://github.com/ax-llm/ax/blob/3d5aa8dac28a2637e4150f022dcd02e6fc3309f0/src/ax/ai/types.ts#L74

serviceTierPricing?

optional serviceTierPricing: Partial<Record<"flex" | "priority", AxServiceTierPricing>>;

Defined in: https://github.com/ax-llm/ax/blob/3d5aa8dac28a2637e4150f022dcd02e6fc3309f0/src/ax/ai/types.ts#L95

Token pricing overrides selected from the tier that actually served a request.

supported?

optional supported: object;

Defined in: https://github.com/ax-llm/ax/blob/3d5aa8dac28a2637e4150f022dcd02e6fc3309f0/src/ax/ai/types.ts#L100

imageOutput?

optional imageOutput: boolean;

Whether this model can return generated images from chat().

operations?

optional operations: readonly ("chat" | "transcribe")[];

Public Ax operations verified for this exact model.

reasoningOffByDefault?

optional reasoningOffByDefault: boolean;

The model does not reason when a request sets no reasoning effort.

samplingWithoutReasoning?

optional samplingWithoutReasoning: boolean;

Reasoning can be turned off (reasoning effort none), and the model then accepts the sampling parameters listed in notSupported.

serviceTiers?

optional serviceTiers: readonly AxServiceTier[];

Portable request tiers verified for this exact model.

showThoughts?

optional showThoughts: boolean;

structuredOutputModes?

optional structuredOutputModes: readonly AxStructuredOutputRung[];

Ordered, verified structured-output strategies for this exact model.

structuredOutputs?

optional structuredOutputs: boolean;

temperatureOne?

optional temperatureOne: boolean;

The model takes an explicit temperature of exactly 1, its default, where notSupported.temperature (or reasoning) rejects other values.

thinkingBudget?

optional thinkingBudget: boolean;