← All projects
API Cost Comparison
Tracking AI model costs across OpenCode Go, DeepSeek Official API, and the OpenRouter marketplace — because you can't optimise what you don't measure.
Output Cost Comparison: Chinese vs American Models
Top 10 most popular models on OpenRouter · Output cost per 1M tokens · Sorted highest → lowest
🇨🇳 Chinese
🇺🇸 American
Pricing at a Glance
OpenCode Go
| Plan | First Month | After | Usage Limits |
|---|---|---|---|
| Go | $5 | $10/mo | $12/5h · $30/wk · $60/mo |
Includes GLM-5.2, Kimi K2.7 Code, DeepSeek V4 Flash/Pro, Qwen3.7 Max, and 8+ other models bundled into the monthly pool. Usage drawn from the $60 monthly cap — cheaper models like V4 Flash get ~31,650 requests per 5h window.
DeepSeek Official API
| Model | Input (cache miss) | Input (cache hit) | Output | Context | Max Output |
|---|---|---|---|---|---|
| deepseek-v4-flash | $0.14 / 1M | $0.0028 / 1M | $0.28 / 1M | 1,048,576 | 384K |
| deepseek-v4-pro | $0.435 / 1M | $0.003625 / 1M | $0.87 / 1M | 1,048,576 | 384K |
Concurrency: 2,500 (Flash) · 500 (Pro). Deprecated aliases deepseek-chat and deepseek-reasoner retire 2026-07-24. V4 Pro's 75% discount is now permanent.
Usage Log (Latest Snapshot)
OpenCode Go
Rolling
2%
resets 3.5h
Weekly
99%
resets 2.5d
Monthly
72%
resets 8d
DeepSeek API (June)
V4 Flash
1,599 req
68.8M tokens
V4 Pro
665 req
47.5M tokens
Balance: $0.00 · June spend: ¥9.52