Supported models and pricing

The Zespan SDK computes cost_usd client-side using a built-in pricing table. This page lists all models currently in that table. If you use a model not listed here, cost_usd will be 0 — see Unsupported models for how to handle this.

Pricing in this table reflects public list prices as of the documentation date. Provider prices change regularly. Zespan updates the SDK pricing table with each release. To get the latest rates, update to the newest SDK version.

OpenAI

Model	Input (per 1M tokens)	Output (per 1M tokens)	Cached input
`gpt-4o`	$2.50	$10.00	$1.25
`gpt-4o-2024-08-06`	$2.50	$10.00	$1.25
`gpt-4o-mini`	$0.15	$0.60	$0.075
`gpt-4o-mini-2024-07-18`	$0.15	$0.60	$0.075
`gpt-4-turbo`	$10.00	$30.00	—
`gpt-4-turbo-2024-04-09`	$10.00	$30.00	—
`gpt-3.5-turbo`	$0.50	$1.50	—
`o1`	$15.00	$60.00	$7.50
`o1-mini`	$3.00	$12.00	$1.50
`o3-mini`	$1.10	$4.40	$0.55
`text-embedding-3-small`	$0.02	—	—
`text-embedding-3-large`	$0.13	—	—
`text-embedding-ada-002`	$0.10	—	—

Anthropic

Model	Input (per 1M tokens)	Output (per 1M tokens)	Cache read	Cache write
`claude-opus-4-7`	$15.00	$75.00	$1.50	$18.75
`claude-sonnet-4-6`	$3.00	$15.00	$0.30	$3.75
`claude-haiku-4-5-20251001`	$0.80	$4.00	$0.08	$1.00
`claude-3-5-sonnet-20241022`	$3.00	$15.00	$0.30	$3.75
`claude-3-5-haiku-20241022`	$0.80	$4.00	$0.08	$1.00
`claude-3-opus-20240229`	$15.00	$75.00	$1.50	$18.75

Google Gemini

Text generation

Model	Input (per 1M tokens)	Output (per 1M tokens)	Cached input
`gemini-2.5-pro`	$1.25	$10.00	$0.31
`gemini-2.5-pro-preview-05-06`	$1.25	$10.00	$0.31
`gemini-2.5-pro-preview-06-05`	$1.25	$10.00	$0.31
`gemini-2.5-flash`	$0.30	$2.50	$0.075
`gemini-2.5-flash-preview-04-17`	$0.15	$0.60	$0.038
`gemini-2.5-flash-lite`	$0.075	$0.30	$0.019
`gemini-2.5-flash-lite-preview-06-17`	$0.075	$0.30	$0.019
`gemini-3.1-flash`	$0.30	$2.50	$0.075
`gemini-3.1-flash-live-preview`	$1.00	$4.00	$0.25
`gemini-2.0-flash`	$0.10	$0.40	$0.025
`gemini-2.0-flash-lite`	$0.075	$0.30	$0.019
`gemini-2.0-pro-exp`	Free	Free	—
`gemini-1.5-pro`	$1.25	$5.00	$0.31
`gemini-1.5-flash`	$0.075	$0.30	$0.019
`gemini-1.5-flash-8b`	$0.038	$0.15	$0.010

Image generation (`span_kind: image_gen`)

Image models return image bytes rather than text tokens. Cost is per image generated and is recorded as input_tokens = image_count with output_tokens = 0.

Model	Price per image
`gemini-3.1-flash-image`	$0.50
`gemini-2.5-flash-image`	$0.30

Text-to-speech

Model	Input (per 1M tokens)	Output (per 1M tokens)	Cached input
`gemini-2.5-flash-tts`	$0.50	$1.00	$0.125
`gemini-2.5-flash-preview-tts`	$0.50	$1.00	$0.125

Video generation — Veo (`span_kind: video_gen`)

Veo models use per-second output pricing. Zespan records cost_usd: 0 and tracks latency of the initiation call. Compute the actual cost from your Google Cloud billing.

Model
`veo-3.1-generate-preview`
`veo-3.1-fast-generate-preview`
`veo-3.1-lite-generate-preview`
`veo-3.0-generate-preview`
`veo-2.0-generate-001`

Embeddings (`span_kind: embedding`)

Model	Input (per 1M tokens)
`gemini-embedding-2`	$0.20
`gemini-embedding-001`	$0.15
`text-embedding-004`	$0.10

AWS Bedrock

Model	Input (per 1M tokens)	Output (per 1M tokens)
`amazon.nova-lite-v1:0`	$0.06	$0.24
`amazon.nova-micro-v1:0`	$0.035	$0.14
`amazon.nova-pro-v1:0`	$0.80	$3.20
`amazon.titan-text-express-v1`	$0.20	$0.60
`amazon.titan-text-lite-v1`	$0.15	$0.20

Mistral

Model	Input (per 1M tokens)	Output (per 1M tokens)
`mistral-large-latest`	$2.00	$6.00
`mistral-small-latest`	$0.20	$0.60
`mistral-7b-instruct`	$0.25	$0.25
`mixtral-8x7b-instruct`	$0.70	$0.70

Groq (hosted models)

Model	Input (per 1M tokens)	Output (per 1M tokens)
`llama-3.3-70b-versatile`	$0.59	$0.79
`llama-3.1-8b-instant`	$0.05	$0.08
`mixtral-8x7b-32768`	$0.24	$0.24
`gemma2-9b-it`	$0.20	$0.20

Unsupported models

If your model is not in the table, the SDK sets cost_usd: 0 and logs a warning at debug level. Your traces will still appear in the dashboard — only the cost column will be blank. To add cost tracking for an unsupported model, you have two options: Option 1 — Compute and pass cost manually:

const response = await myCustomModel.complete(prompt);

const { span } = startSpan({
  name: "custom-model-call",
  model: "my-custom-model",
  provider: "custom",
});

await span.end({
  status: "success",
  input_tokens: response.usage.input,
  output_tokens: response.usage.output,
  cost_usd: (response.usage.input * 0.002 + response.usage.output * 0.006) / 1000,
});

Option 2 — Request model support: Open an issue on GitHub with the model name and public pricing URL. New models are typically added within one release cycle.

Cached token pricing

Cached tokens are billed at a reduced rate when the model provider serves them from its prompt cache. Zespan extracts cached token counts automatically:

OpenAI: response.usage.prompt_tokens_details.cached_tokens
Anthropic: response.usage.cache_read_input_tokens
Google: response.usageMetadata.cachedContentTokenCount

The cost formula used by the SDK:

cost_usd = (input_tokens × input_price
           + output_tokens × output_price
           + cached_tokens × cached_price) ÷ 1,000,000

Tracing & Observability

Evaluations

Datasets & Experiments

Guardrails

Prompt Management

Alerts & Incidents

Cost & Analytics

ZespanPilot (AI Copilot)

Supported models and pricing

OpenAI

Anthropic

Google Gemini

Text generation

Image generation (`span_kind: image_gen`)

Text-to-speech

Video generation — Veo (`span_kind: video_gen`)

Embeddings (`span_kind: embedding`)

AWS Bedrock

Mistral

Groq (hosted models)

Unsupported models

Cached token pricing

​OpenAI

​Anthropic

​Google Gemini

​Text generation

​Image generation (span_kind: image_gen)

​Text-to-speech

​Video generation — Veo (span_kind: video_gen)

​Embeddings (span_kind: embedding)

​AWS Bedrock

​Mistral

​Groq (hosted models)

​Unsupported models

​Cached token pricing

OpenAI

Anthropic

Google Gemini

Text generation

Image generation (`span_kind: image_gen`)

Text-to-speech

Video generation — Veo (`span_kind: video_gen`)

Embeddings (`span_kind: embedding`)

AWS Bedrock

Mistral

Groq (hosted models)

Unsupported models

Cached token pricing