Same models. Smaller numbers.
CheapRelay resells Claude, GPT, Gemini and a dozen other model families by the token. The bars show what the vendor lists against what you pay here.
Every public model, cheapest first
Ordered by input plus output price. Four columns because four kinds of token are priced separately.
| Model | Who makes it | Input | Output | Write to cache | Read from cache |
|---|---|---|---|---|---|
| Qwen3.7 Plus qwen3.7-plus | Alibaba/Qwen | $0.1208 | $0.3625 | $0.00 | $0.0361 |
| MiMo V2.6 Pro MiMo-V2.6-Pro | Xiaomi/MiMo | $0.1736 | $0.3472 | $0.00 | $0.002778 |
| MiniMax M3 MiniMax-M3 | MiniMax | $0.1236 | $0.4917 | $0.00 | $0.025 |
| GPT-6.1 Sol gpt-6.1-sol | OpenAI | $0.1111 | $0.5528 | $0.00 | $0.0264 |
| DeepSeek V4 Flash deepseek-v4-flash | DeepSeek | $0.2389 | $0.9556 | $0.00 | $0.005556 |
| Hunyuan 4 hy4-preview | Tencent/Hunyuan | $0.2861 | $0.9139 | $0.00 | $0.0139 |
| Gemini 2.5 Flash gemini-2.5-flash | $0.1778 | $1.42 | $0.1778 | $0.0181 | |
| Claude Haiku 4.5 claude-haiku-4-5-20251001 | Anthropic | $0.2917 | $1.40 | $0.35 | $0.025 |
| GLM-5.3 glm-5.3 | Zhipu/GLM | $0.4583 | $1.38 | $0.00 | $0.1153 |
| Grok 4.6 grok-4.6 | xAI/Grok | $0.4653 | $1.39 | $0.00 | $0.1167 |
| GPT-5.6 Sol gpt-5.6-sol | OpenAI | $0.3056 | $1.71 | $0.00 | $0.0264 |
| Grok 4.7 grok-4.7 | xAI/Grok | $0.5056 | $1.52 | $0.00 | $0.1264 |
| Qwen3.6 Plus Uncensored qwen3.6-plus-uncensored | Qwen/uncensored | $0.3431 | $1.83 | $0.00 | $0.00 |
| Qwen3.8 Max qwen3.8-max | Alibaba/Qwen | $0.5722 | $1.72 | $0.00 | $0.0681 |
| GPT-6 Astra gpt-6-astra | OpenAI | $0.5639 | $2.82 | $0.00 | $0.0569 |
| DeepSeek V4 Pro deepseek-v4-pro | DeepSeek | $1.01 | $3.03 | $0.00 | $0.0333 |
| Gemini 3.8 Flash gemini-3.8-flash | $0.95 | $4.28 | $0.95 | $0.0944 | |
| Claude Sonnet 5.5 claude-sonnet-5-5 | Anthropic | $0.8903 | $4.45 | $1.11 | $0.0778 |
| Kimi K3 kimi-k3 | Moonshot/Kimi | $0.9333 | $4.67 | $0.00 | $0.0931 |
| Claude Sonnet 5 claude-sonnet-5 | Anthropic | $0.9764 | $4.88 | $1.22 | $0.0861 |
| Claude Opus 5.5 claude-opus-5-5 | Anthropic | $1.32 | $6.24 | $1.56 | $0.0597 |
| Gemini 3.1 Pro gemini-3.1-pro | $1.12 | $6.49 | $1.12 | $0.1125 | |
| Claude Opus 5 claude-opus-5 | Anthropic | $1.66 | $8.28 | $2.04 | $0.1222 |
All prices per 1,000,000 tokens. Chinese pages show yuan; other languages show the US-dollar equivalent.
Why it's cheap, honestly
You pay per token actually used: input, output, cache writes and cache reads each at their own rate. No monthly fee, no seat licence, no minimum spend.
We own no GPUs. Calls travel through a wholesale upstream that several relays share, and we resell that capacity with a thin margin on top.
No sales team, no account managers, no glossy uptime promises. Support works through dashboard messages and email, answered by the people who run the relay.
What you will not find: subscription traps, hidden surcharges, or charges for calls the upstream rejected. Every call is itemised in your dashboard with its tokens and price.