Agents
Tool-calling loops where calls-per-task is high and unit price compounds.
Model catalog · Managed AI Gateway
MiniMax fits agent loops and coding tasks — tool calls and structured output without stepping up to frontier-model pricing.
AI24X flash $0.35/M output · pro $1.05/M output · flat, verified 2026-10-02
Tool-calling loops where calls-per-task is high and unit price compounds.
Generation and refactoring at flash-tier cost.
Members can request MiniMax by name.
Named models & access →Base URL: https://api.ai24x.com/v1 · Example model ID: vip-minimax
curl https://api.ai24x.com/v1/chat/completions \
-H "Authorization: Bearer $AI24X_API_KEY" \
-H "Content-Type: application/json" \
-d '{"model":"vip-minimax","messages":[{"role":"user","content":"Hello"}],"max_tokens":200}'
flash / pro for first tests. Exact model IDs and access rules are shown in your account and the live model catalog.One API key covers DeepSeek, Qwen, GLM, Kimi, MiniMax, MiMo and frontier models — on a flat rate with no peak-hour surcharge and no request-size price tiers. Payment is by PayPal or card, and no Chinese phone number is required.
Can I use the MiniMax API outside China without a Chinese phone number?
Yes. AI24X provides OpenAI-compatible access through one endpoint (https://api.ai24x.com/v1). Signup does not require a Chinese phone number and payment is accepted by PayPal or card.
Does AI24X charge a peak-hour surcharge?
No. AI24X bills a flat rate with no peak/off-peak window surcharge and no request-size price tiers.
Is AI24X the official MiniMax API?
No. AI24X is an independent AI gateway and is not affiliated with any model provider. Model names describe compatibility only.