Redpill
OpenAI-compatible confidential inference on Phala GPU TEE
Redpill by Phala Network offers confidential AI inference running on Phala's decentralized GPU TEE infrastructure. Models execute inside Intel TDX enclaves with NVIDIA H100 Confidential Computing, providing cryptographic proof that inference is private. Fully OpenAI-compatible API with transparent on-chain attestation.
API Features
OpenAI Compatible ✓
Streaming ✓
Function Calling ✓
Batch Inference —
Vision ✓
Embeddings ✓
Models & Pricing
| Model | Family | Context | Max Out | Input /M | Output /M | Free |
|---|---|---|---|---|---|---|
| GPT-OSS 20B | OpenAI GPT | 128K | — | $0.04 | $0.15 | — |
| GPT-OSS 120B | OpenAI GPT | 128K | — | $0.10 | $0.49 | — |
| Gemma 3 27B | Google Gemma | 52K | — | $0.11 | $0.40 | — |
| Qwen2.5 7B | Qwen | 32K | — | $0.04 | $0.10 | — |
| DeepSeek V3.2 | DeepSeek | 160K | — | $0.27 | $0.40 | — |
| Qwen3 VL 30B | Qwen | 128K | — | $0.20 | $0.70 | — |
| Qwen2.5 VL 72B | Qwen | 64K | — | $0.20 | $0.70 | — |
| Venice Uncensored 24B | Custom | 32K | — | $0.20 | $0.90 | — |
| GLM-4.7 Flash | ZhipuAI GLM | 198K | — | $0.10 | $0.43 | — |
| Kimi K2.5 | Moonshot AI | 256K | — | $0.60 | $3.00 | — |
| GLM-5 | ZhipuAI GLM | 198K | — | $1.20 | $3.50 | — |