Loading template…

Calling GPT-6.1, Claude and Gemini APIs from mainland China

Use a ModelFold API key and the OpenAI-compatible base URL https://modelfold.ai/v1 to integrate supported GPT-6.1, Claude and Gemini models. Retrieve the exact API model ID before sending a request; curl and Python examples are below. Mainland connectivity varies by network and model availability, and is not an uptime guarantee.

Official sources checked:

Price snapshot updated:

Related guides

Current public prices

Pay as you go with prepaid credits. 200 credits = US$1. Credit charges round up to two decimal places; this is not USD rounding. Failed generations are not charged; a temporary hold may be released after failure. Review the final quote before submitting. Token rates are per one million tokens, input / output; video prices are starting prices for the listed configuration.

ModelModelFold USDOfficial USDUnit / configuration
Claude Haiku 5.50.072 in / 0.36 out$0.1 / $0.5 (≤100,000 input tokens)per 1M tokens
Claude Opus 52.40 in / 12.00 outNot independently verifiedper 1M tokens
Claude Opus 5.51.92 in / 9.60 out$4 / $20per 1M tokens
Claude Sonnet 51.02 in / 5.13 outNot independently verifiedper 1M tokens
Claude Sonnet 5.50.96 in / 4.80 out$2 / $10per 1M tokens
Gemini 3 Pro0.60 in / 4.20 outNot independently verifiedper 1M tokens
Gemini 3.1 Pro0.60 in / 4.20 outNot independently verifiedper 1M tokens
Gemini 3.6 Flash0.27 in / 1.35 outNot independently verifiedper 1M tokens
Gemini 3.7 Flash0.27 in / 1.35 outNot independently verifiedper 1M tokens
Gemini 3.8 Flash0.27 in / 1.35 outNot independently verifiedper 1M tokens
Gemini 4 Argon0.96 in / 4.80 outNot independently verifiedper 1M tokens
Gemini Omni 1.1 Flash0.024 in / 0.024 outNot independently verifiedper 1M tokens
GPT 6.1 Sol0.72 in / 3.60 out$2 / $10per 1M tokens
GPT-6 Astra3.36 in / 16.80 out$10 / $50per 1M tokens
GPT-6 Sol and Luna0.036 in / 0.18 out$0.1 / $0.5per 1M tokens
Gemini 4 Carbon2.40 in / 12.00 outNot independently verifiedper 1M tokens
GPT-6.1 Terra1.20 in / 6.00 out$1 / $5per 1M tokens
Gemini 3.1 Pro Preview2.40 in / 14.40 outNot independently verifiedper 1M tokens
Gemini 3.5 Flash Lite0.36 in / 3.00 outNot independently verifiedper 1M tokens
Gemini 3.1 Flash Lite0.30 in / 1.80 outNot independently verifiedper 1M tokens
Gemini 3.1 Flash Lite Preview0.30 in / 1.80 outNot independently verifiedper 1M tokens
Gemini 3 Flash Preview0.60 in / 3.60 outNot independently verifiedper 1M tokens
Gemini 2.5 Flash0.36 in / 3.00 outNot independently verifiedper 1M tokens
Gemini 2.5 Flash Lite0.12 in / 0.48 outNot independently verifiedper 1M tokens
GPT-6 Astra Pro12.00 in / 60.00 outNot independently verifiedper 1M tokens
GPT-6 Luna Pro0.12 in / 0.60 outNot independently verifiedper 1M tokens
Gemini 3.1 Pro Preview Custom Tools2.40 in / 14.40 outNot independently verifiedper 1M tokens
Gemini 4 Hydrogen1.80 in / 9.00 outNot independently verifiedper 1M tokens
GPT-6.1 Terra Pro1.20 in / 6.00 outNot independently verifiedper 1M tokens
GPT-6.1 Sol Pro2.40 in / 12.00 outNot independently verifiedper 1M tokens
GPT-6 Sol Pro2.40 in / 12.00 outNot independently verifiedper 1M tokens

API examples

curl https://modelfold.ai/v1/models -H "Authorization: Bearer $MODELFOLD_API_KEY"
curl https://modelfold.ai/v1/chat/completions \
  -H "Authorization: Bearer $MODELFOLD_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{"model":"claude-haiku-5-5","messages":[{"role":"user","content":"Hello"}]}'
import os
from openai import OpenAI

client = OpenAI(api_key=os.environ["MODELFOLD_API_KEY"], base_url="https://modelfold.ai/v1")
response = client.chat.completions.create(
    model="claude-haiku-5-5",
    messages=[{"role": "user", "content": "Hello"}],
)
print(response.choices[0].message.content)

Examples are illustrative and were not sent to a paid model. Install the OpenAI Python SDK separately. Video requests are asynchronous; follow the model documentation for polling and required parameters.

Frequently asked questions

Is ModelFold always the cheapest?

No. Compare the same model, context, cache mode and input/output mix. Final quotes and actual usage determine the charge.

Are failed generations charged?

Failed generations are not charged. A temporary reservation is not a final charge; check account logs for settlement.

Where do I find the correct model ID?

Use GET /v1/models and the linked model documentation. A website page slug is not always the API model ID.

Comparison sources