Per-token, per-character, per-second. No subscription, no minimums.
All prices already include RelayX margin (10% on LLM/embeddings/ASR, 5% on TTS).
| Model | Input | Output | Cache read |
|---|---|---|---|
| deepseek/deepseek-v4-flash | $0.17/ 1M tokens | $0.34/ 1M tokens | $0.003/ 1M tokens |
| deepseek/deepseek-v4-pro | $0.53/ 1M tokens | $1.05/ 1M tokens | $0.005/ 1M tokens |
| anthropic/claude-4-7-sonnet | $3.30/ 1M tokens | $16.50/ 1M tokens | $0.330/ 1M tokens |
| anthropic/claude-4-7-opus | $5.50/ 1M tokens | $27.50/ 1M tokens | $0.550/ 1M tokens |
| openai/gpt-5 | $1.38/ 1M tokens | $11.00/ 1M tokens | $0.138/ 1M tokens |
| openai/gpt-5-mini | $0.28/ 1M tokens | $2.20/ 1M tokens | $0.028/ 1M tokens |
| google/gemini-3-pro | $1.38/ 1M tokens | $11.00/ 1M tokens | — |
| qwen/qwen3-max | $0.42/ 1M tokens | $1.67/ 1M tokens | — |
| openai/gpt-4.1 | $2.20/ 1M tokens | $8.80/ 1M tokens | — |
| openai/gpt-4.1-mini | $0.44/ 1M tokens | $1.76/ 1M tokens | — |
| openai/gpt-4.1-nano | $0.11/ 1M tokens | $0.44/ 1M tokens | — |
| openai/gpt-5-nano | $0.06/ 1M tokens | $0.44/ 1M tokens | — |
| openai/gpt-5-codex | $1.38/ 1M tokens | $11.00/ 1M tokens | — |
| openai/o4-mini | $1.21/ 1M tokens | $4.84/ 1M tokens | — |
| google/gemini-3-flash | $0.33/ 1M tokens | $2.75/ 1M tokens | — |
| google/gemini-3-flash-lite | $0.28/ 1M tokens | $1.65/ 1M tokens | — |
| xai/grok-4-fast | $0.22/ 1M tokens | $0.55/ 1M tokens | — |
| moonshot/kimi-k2 | $0.61/ 1M tokens | $2.42/ 1M tokens | — |
| zai/glm-4-6 | $0.07/ 1M tokens | $0.24/ 1M tokens | $0.012/ 1M tokens |
| alibaba/qwen3-coder-plus | $0.59/ 1M tokens | $2.38/ 1M tokens | $0.119/ 1M tokens |
| qwen/qwen3-vl-flash | $0.02/ 1M tokens | $0.23/ 1M tokens | $0.005/ 1M tokens |
| qwen/qwen3-vl-plus | $0.15/ 1M tokens | $1.51/ 1M tokens | $0.030/ 1M tokens |
| minimax/m2 | $0.11/ 1M tokens | $0.11/ 1M tokens | — |
| Model | Description | Input |
|---|---|---|
| dashscope/text-embedding-v4 | Qwen3-Embedding family — Chinese-first dense text embeddings | $0.077/ 1M tokens |
| Model | Per 1M chars |
|---|---|
| vox/index-tts-2 | $1.575/ 1M chars |
| vox/voice-clone-v1 | $1.575/ 1M chars |
| minimax/speech-2.6-hd | $105.000/ 1M chars |
| minimax/speech-2.6-turbo | $63.000/ 1M chars |
| Model | Per minute |
|---|---|
| alibaba/paraformer-v2 | $0.0319/ minute |
| Model | Description | Per image |
|---|---|---|
| rx-image-flux | Flux — high quality, balanced speed | $0.003/ image |
| rx-image-qwen | Qwen-image — strong on Chinese prompts | $0.002/ image |
| rx-image-z | Z-image — premium aesthetic | $0.003/ image |
| rx-image-z-pro | Z-image (DashScope direct) — reliable, low-latency premium tier | $0.014/ image |
| rx-image-qwen-edit | Qwen-Image-Edit — reference-image conditioned generation (0–3 refs) | $0.014/ image |
| rx-image-gpt-draft | GPT Image (draft) — cheapest tier, for composition search | $0.016/ image |
| rx-image-gpt | GPT Image — strong prompt adherence, text-to-image or 0–3 refs | $0.063/ image |
| rx-image-gpt-pro | GPT Image (high) — maximum fidelity, for final renders | $0.250/ image |
| Model | Description | Per second |
|---|---|---|
| rx-video-ltx | LTX-2.3 — fastest image-to-video, 720p-friendly | $0.020/ second |
| rx-video-wan | Wan2.2-Lightning — higher-quality image-to-video | $0.030/ second |
| rx-video-h3 | MiniMax H3 — text/reference to video with native stereo audio | $0.050/ second |
| rx-video-animate | Wan 2.2 Animate — copy a driving video's motion onto a character image | $0.040/ second |
| rx-avatar-talk | InfiniteTalk — portrait + speech to a lip-synced talking video (length follows the audio) | $0.015/ second |
| rx-avatar-dub | InfiniteTalk — re-voice an existing clip, keeping its framing and camera movement | $0.012/ second |
Contact us on WeChat for now (Stripe coming once we incorporate).
Each request is charged to your balance at the rates above.
Per-call usage + cost is itemized in real time.