99.9% SLANo Read me document is available for this model yet.
Claude Haiku 5.5
Claude Haiku 5.5
- 99.9% SLA
- Official Discounts
- Pay-as-you-go
- Fast & Low Latency
Brand:
Price: Tiered pricing
Price: Tiered pricingCompatible with common protocols
Integrate using the protocol your team already knows. All examples below keep the vendor's official Python SDK and only replace the gateway base URL with the MuseSpark-compatible endpoint.
from openai import OpenAI
client = OpenAI(
api_key="YOUR_API_KEY",
base_url="https://api.musespark.ai/proxy/openai/v1",
)
response = client.responses.create(
model="claude-haiku-5.5",
input="Write a one-sentence release note for a new image model.",
)
print(response.output_text)Protocol reference
All three routes reach the same model service. The main difference is the request format and client calling style.
OpenAI-compatible
Responses API shape
Designed for teams already using the OpenAI SDK and response-style objects.
Endpoint
POST https://api.musespark.ai/proxy/openai/v1/responses
Auth
Authorization: Bearer <API_KEY>
Read response
Use response.output_text first, then inspect output[] when structured items matter.
Gemini-compatible
generateContent API shape
A natural fit for apps already built around contents[] and parts[] prompt builders.
Endpoint
POST https://api.musespark.ai/proxy/gemini/v1beta/models/claude-haiku-5.5:generateContent
Auth
Authorization: Bearer <API_KEY> or X-API-Key: <API_KEY>
Read response
Merge candidates[0].content.parts[] or use the text convenience field when available.
Anthropic-compatible
Messages API shape
For Claude-style chat flows that already think in messages[] and content blocks.
Endpoint
POST https://api.musespark.ai/proxy/anthropic/v1/messages
Auth
Authorization: Bearer <API_KEY>
Read response
Extract text blocks from content[] when you want the final answer string.
SDK installation
Install the official SDK that matches your integration style, then point requests to the corresponding MuseSpark-compatible endpoint.
Official OpenAI Python SDK
Install the official OpenAI SDK
Use OpenAI's official Python SDK when your team wants to call the MuseSpark gateway through the OpenAI-compatible client surface.
pip install openaiOfficial package: openai. The MuseSpark-specific step is setting base_url to /proxy/openai/v1.
Official Google Gen AI SDK
Install the official Google Gen AI SDK
Use Google's official Gemini SDK for Python when your existing app already works with contents, parts, and generate_content calls.
pip install google-genaiOfficial package: google-genai. The MuseSpark-specific step is pointing http_options.base_url at /proxy/gemini.
Official Anthropic Python SDK
Install the official Anthropic SDK
Use Anthropic's official Python SDK when your application already wraps Claude-style message flows and wants to keep the same client ergonomics.
pip install anthropicOfficial package: anthropic. The MuseSpark-specific step is setting base_url to /proxy/anthropic.
Why Choose MuseSpark?
Token billing · postpaid
Billed according to actual usage. Final charges are based on actual token usage.
Rates are USD / 1M tokens (coefficients); the first matching tier is used; if none match, the price is unavailable. Listed adjustments and formulas apply to the coefficients; effective rates are shown where they can be resolved without request-specific values.
Input tokens ≥ 0
Output tokens ≥ 0
cache_read ≥ 0
cache_write ≥ 0
total_input_tokens ≥ 0
total_input_tokens ≤ 100000
- Input token
- $0.10/M
- Output token
- $0.50/M
- Input cache read
- $0.01/M
- Input cache write
- $0.125/M
- Image input token
- $0.00/M
- Image output token
- $0.00/M
Pricing formula
(max(input - cache_read, 0) * input_factor_per_1m + output * output_factor_per_1m + cache_read * read_cache_factor_per_1m + cache_write * create_cache_factor_per_1m) / 1000000
Input tokens ≥ 0
Output tokens ≥ 0
cache_read ≥ 0
cache_write ≥ 0
total_input_tokens > 100000
- Input token
- $0.50/M effective · base $0.10/M
- Output token
- $2.50/M effective · base $0.50/M
- Input cache read
- $0.05/M effective · base $0.01/M
- Input cache write
- $0.625/M effective · base $0.125/M
- Image input token
- $0.00/M
- Image output token
- $0.00/M
Multiplier ×5
Pricing formula
(max(input - cache_read, 0) * input_factor_per_1m + output * output_factor_per_1m + cache_read * read_cache_factor_per_1m + cache_write * create_cache_factor_per_1m) / 1000000
Examples
No examples yet.
FAQ
No FAQs are available for this item yet.
