OpenCode
OpenCode Go
Subscription sticker price for OpenCode · surfaces: coding. List rates below are API prices for the same models, not what the seat costs.
Low-cost agentic coding subscription from OpenCode (Anomaly). $5 first month, then $10/mo; cancel any time. Works with OpenCode or any agent via the Go API key. Included usage is dollar-budgeted (~$12 per 5h, $30/week, $60/month), so request counts depend on model cost. Vendor estimates at typical Go patterns: Grok 4.5 ~120 req/5h; Kimi K3 ~110; Qwen3.8 Max ~160; GLM-5.2 ~880; MiniMax M3 ~3,200; DeepSeek V4 Pro ~3,450; GPT-5.6 Luna ~2,050 (promo 2× usage may apply); DeepSeek V4 Flash ~31,650; MiMo-V2.5 ~30,100. Some models get a lower usage multiplier (~$15 of included usage) until bulk rates land. After limits, optional Zen balance top-up or free OpenCode models. One Go subscriber per workspace. Model roster changes as OpenCode retests providers. Source: https://opencode.ai/go and https://opencode.ai/docs/go/
Get $5 usage credit toward your Go usage limits
Referral link. The credit lands on your side at sign-up, and we get one too, which goes back into developing this site.
Models on this plan
Representative lineup. $/MTok columns are published API list rates for comparison, not included subscription value. Checked . 16 with benchmark scores.
| Model | Tier | In $/M | Out $/M | Context | Score |
|---|---|---|---|---|---|
Grok 4.5xai/grok-4.5 | frontier | — | — | 1M | 1549Arena code |
Kimi K3moonshotai/kimi-k3 | frontier | $2.70 | $13.50 | 1.0M | 9.91Arena agent |
Qwen3.8 Maxqwen/qwen3.8-max | frontier | $2 | $6 | 1M | 1668Arena code |
GLM 5.2zai/glm-5.2 | strong | — | — | 1.0M | 51.1AA intel |
GLM-5.1zai/glm-5.1 | strong | $1.40 | $4.40 | 202K | 40.2AA intel |
GPT-5.6 Lunaopenai/gpt-5.6-luna | strong | — | — | 1.1M | 51.2AA intel |
Kimi K2.7 Codemoonshotai/kimi-k2.7-code | strong | $0.71 | $3.50 | 262K | 41.9AA intel |
Kimi K2.6moonshot/kimi-k2.6 | strong | $0.95 | $4 | 262K | 44.2AA intel |
MiniMax M3minimax/minimax-m3 | strong | — | — | 1.0M | 44.4AA intel |
MiniMax M2.7MiniMax/MiniMax-M2.7 | strong | — | — | 205K | 38.1AA intel |
DeepSeek V4 Prodeepseek/deepseek-v4-pro | strong | $0.435 | $0.87 | 1.0M | 44.3AA intel |
Qwen3.7 Maxqwen/qwen3.7-max | strong | — | — | 1M | 46AA intel |
Qwen3.7 Plusqwen/qwen3.7-plus | fast | — | — | 1M | 39AA intel |
Qwen3.6 Plusqwen/qwen3.6-plus | fast | — | — | 1M | 39.6AA intel |
DeepSeek V4 Flash 0423deepseek/deepseek-v4-flash | fast | $0.14 | $0.28 | 1.0M | — |
MiMo V2.5xiaomi/mimo-v2.5 | fast | — | — | 1.1M | 37.2AA intel |
MiMo-V2.5-Procrof/mimo-v2.5-pro | strong | — | — | 1.1M | 42.2AA intel |
Hy3cortecs/hy3 | fast | — | — | 262K | — |
Explore around this plan
- All OpenCode subscription plansSticker ladder for this vendor
- OpenCode in the provider indexHosts, gateways, and model counts
- Full models catalogList rates, context, scores across providers
- Compare the first two linked modelsSide-by-side rates and scores
- Usage calculatorShape input/output/cache against list rates
- Vendor source pageAuthoritative sticker price and terms
Score sources on /benchmarks.
See what your real usage is worth
Sticker prices are not usage value at list rates. tokenstat measures the sessions on your machine so you can see the scale of the work, not a fake break-even.
Looking for tokenstat Free / Supporter / Patron? That is /pricing, not this page.