AMD Radeon Cloud API Key Guide: Free Access to DeepSeek V4 Flash, Qwen 3.8, and 4 Flagship Models

AMD Radeon Cloud API Key Guide: Free Access to DeepSeek V4 Flash, Qwen 3.8, and 4 Flagship Models

AMD China Developer Platform has launched Token Factory on Radeon Cloud, offering free public API access to 4 flagship models: DeepSeek-V4-Flash (1M context), DeepSeek-V4-Flash-Vision-Exp, Qwen 3.8 Flash Next, and MiniCPM 5 1B.

Real developer testing confirms AMD's exceptional availability—unlike NVIDIA NIM which frequently encounters strict 429 limits, AMD's serverless endpoints respond reliably. With native dual OpenAI & Anthropic protocol compatibility, OpenCode works right out of the box by simply adding your key.

💡 Key Highlights

• 🌐 Base URL: https://developer.amd.com.cn/radeon/api/v1 (OpenAI: /chat/completions, Anthropic: /messages)
• 🤖 4 Free Models: DeepSeek-V4-Flash (1M), DeepSeek-V4-Flash-Vision-Exp (1M), Qwen3.8-Flash-Next (256K), MiniCPM5-1B (128K)
• 🚀 OpenCode Integration: Add your key directly in OpenCode. (Note: ZCode does not currently support this endpoint).
• 🎁 Billing: Public Free Model APIs with zero credit card requirements. Points indicate relative volume, not charges.

🔑 3 Steps to Retrieve Your API Key

1. Sign in at https://developer.amd.com.cn/radeon/tokenfactory.
2. Click any free model card (e.g. DeepSeek-V4-Flash-0731) and click Copy next to API Key.
3. In OpenCode, set Base URL to https://developer.amd.com.cn/radeon/api/v1 and paste your key.