Skip to main content
A2Agent makes it simple to access China’s leading large language models from anywhere outside mainland China — at a fraction of official pricing, with no monthly fees and no lock-in. Below you’ll find answers to the questions we hear most often, organized by topic.
Default model groups connect directly to the official provider APIs. A2Agent never downgrades a model or tampers with request or response content. If you need a lower-cost option, lower-priced groups are available and are clearly labeled as such on the group page, so you always know exactly what you’re using.
For coding, we recommend the following models based on benchmark performance and user feedback:
  • deepseek-v4-pro — strong reasoning and multi-language code generation
  • glm-5.1 — excellent instruction following and code completion
  • kimi-k2.7-code — purpose-built for code with a large context window
  • minimax-m2.5 — solid all-rounder for code review and explanation
You can compare model IDs and pricing on the Models Overview page before committing to a specific model.
Yes. kimi-k2.5 and kimi-k2.6 both accept vision inputs, allowing you to pass images alongside text prompts. Pass image content as a base64-encoded data URL or a publicly accessible HTTPS URL in the content array of your message, following the standard OpenAI-compatible multimodal message format.
Context window sizes vary by model family:
Model familyMax context
Kimi256K tokens
GLM203–205K tokens
MiniMax200K tokens
DeepSeek128K tokens
Always check the specific model’s detail page for the authoritative context limit, as it can vary between versions within the same family.
No. Your A2Agent balance never expires. Credits sit in your account until you spend them — there are no monthly rollovers, no forfeiture windows, and no inactivity fees. Your balance is only deducted when you make API calls.
A2Agent accepts the following payment methods:
  • USDT (on-chain crypto)
  • Stripe (card processing)
  • PayPal
  • Overseas credit and debit cards
All payments are processed in USD. Mainland China payment methods are not supported.
No. A2Agent is pure pay-as-you-go. You top up your balance and pay only for the tokens you consume. There is no base fee, no seat fee, and no minimum monthly spend.
Free trial credits are automatically added to your account when you complete registration. Sign up at https://a2agent.me/register — no credit card is required to claim your trial credits.
Trial credits are a great way to benchmark models and measure costs before committing to a top-up.
No. A2Agent is designed exclusively for users and entities outside mainland China. Before registering or making API calls, please confirm that your location, account entity, and intended use case comply with your local laws and the terms of service of the underlying model providers. Accounts found to be operating from mainland China may be suspended without notice.
Any region outside mainland China is supported. There are no other geographic restrictions. Payments can be made using overseas cards and the payment methods listed above. If you’re unsure whether your region qualifies, contact us at patchx2025@gmail.com before signing up.
A2Agent does not persist your request payloads. Your prompts and model responses are forwarded to provider APIs and are never written to long-term storage by A2Agent. Only the metadata required for metering, billing, and risk control — such as timestamps, token counts, and request rates — is retained. Your inputs and outputs are never used for model training or shared with third parties. See the Privacy Policy page for the full breakdown.
Your raw API key is displayed only once at creation time. After that, only a hashed representation is stored. To keep your key safe:
  • Never share it publicly or via chat
  • Never commit it to source control (use environment variables instead)
  • Revoke and regenerate immediately from the console if you suspect it has been exposed
If you lose your key, you must generate a new one — the original cannot be recovered.
A 401 response means your request failed authentication. Check the following:
  1. Your Authorization header is formatted as Bearer YOUR_KEY (not just the raw key).
  2. The base URL is exactly https://api.a2agent.me — no trailing slash, no extra path prefix.
  3. The key hasn’t been revoked. Generate a new key from the console if needed.
  4. You’re not accidentally using a key from a different account or provider.
A 403 response means your request was rejected after authentication. There are two common causes:
  • Suspended account — your account may have been suspended. Contact patchx2025@gmail.com to find out why and resolve it.
  • Insufficient balance — your account balance has been exhausted. Top up your account from the console to resume making API calls.
A 403 is distinct from a 401: a 401 means the key itself is invalid or missing, while a 403 means the key is valid but the account cannot currently fulfill requests.
A 429 response means you’ve exceeded the rate limit for your current tier. To resolve this:
  • Implement exponential back-off and retry logic in your client.
  • Reduce request concurrency if you’re firing many parallel calls.
  • Contact support at patchx2025@gmail.com or join the Discord to discuss higher rate-limit tiers.
If a model’s output seems wrong or inconsistent with what you expect, start here:
  1. Verify the model ID is spelled exactly as shown on the Models Overview page — IDs are case-sensitive.
  2. Check that you’re not accidentally hitting a lower-priced group when you intend to use the default group.
  3. Review your system prompt and message structure for unintended instructions.
  4. If the issue persists, share the request details (minus sensitive content) in the Discord for community help.
Still have questions? Reach out at patchx2025@gmail.com or join the community on Discord.