OpenCode Go: limits, models and real cost
OpenCode Go costs $10 a month. It runs Kimi K3.1 Pro, Kimi K3, GLM 5.2, GLM 5.3, Qwen3.8 Max and 18 more. OpenCode describes the allowance as “Curated, affordable models; per-model requests per 5 hours and monthly usage published”. How it refreshes: 5-hour and monthly limits. Used in full on DeepSeek V4.1 Flash, the allowance comes to about 10 billion tokens a month by Real API Pricing's medium-confidence estimate: about $170 at API prices, 17× the subscription.
- Price
- $10
- a month
- Allowance in dollars
- $170
- estimate on DeepSeek V4.1 Flash
- API usage per dollar
- 17×
- at full use
- Break-even
- 42 sessions
- a month on Kimi K3
About OpenCode
Three ways to pay, only one of them a subscription. The editor itself is free with your own provider key; Console bills per request from prepaid credits; Go and Go Plus are flat monthly plans over a fixed list of mostly open models, and the Go page publishes estimated requests per five hours and a monthly dollar budget for each model.
Limits and reset windows
| In OpenCode’s words | “Curated, affordable models; per-model requests per 5 hours and monthly usage published” |
|---|---|
| How it refreshes | 5-hour and monthly limits |
| How the size is stated | Not published |
Models, effort and what a task costs
One task of Artificial Analysis' Intelligence Index on Kimi K3.1 Pro at max effort costs $3.68 through the API and about $2.46 through OpenCode Go at full use. In T3 Code's usage data an average Claude Haiku 5.5 turn uses about 1.7 million tokens, 96% of them read back from cache, and costs $0.030 at API list price, so a full allowance holds roughly 494 such turns a month.
| Model | Effort | Index | One task, API | One task, plan | Tasks a month |
|---|---|---|---|---|---|
| Kimi K3.1 Pro | max | 59.9 | $3.68 | $2.46 | 4 |
| GLM 5.3 | low | 47.3 | $1.57 | $1.05 | 10 |
| GLM 5.3 | max | 58.8 | $5.01 | $3.34 | 3 |
| Qwen3.8 Max | max | 58.7 | $10.02 | $6.68 | 1 |
| Grok 4.7 | low | 55.4 | $2.92 | $1.95 | 5 |
| Grok 4.7 | high | 56.9 | $6.18 | $4.13 | 2 |
| Grok 4.7 | xhigh | 57.4 | $7.77 | $5.19 | 2 |
| Claude Haiku 5.5 | low | 47.7 | $0.044 | $0.029 | 343 |
| Claude Haiku 5.5 | medium | 50.1 | $0.092 | $0.061 | 163 |
| Claude Haiku 5.5 | high | 55.3 | $0.16 | $0.11 | 95 |
| Claude Haiku 5.5 | xhigh | 56.7 | $0.25 | $0.17 | 60 |
| Claude Haiku 5.5 | max | 55.3 | $0.41 | $0.27 | 36 |
| GPT-5.6 Luna | max | 56.2 | $0.33 | $0.22 | 45 |
| DeepSeek V4 Pro | high | 56.0 | $0.85 | $0.36 | 27 |
| DeepSeek V4 Pro | max | 54.7 | $1.77 | $0.76 | 13 |
| GPT-6 Luna | low | 45.8 | $0.0051 | $0.0034 | 2,930 |
| GPT-6 Luna | medium | 49.6 | $0.042 | $0.028 | 355 |
| GPT-6 Luna | high | 51.1 | $0.073 | $0.049 | 206 |
| GPT-6 Luna | xhigh | 52.0 | $0.10 | $0.069 | 144 |
| GPT-6 Luna | max | 55.1 | $0.15 | $0.10 | 99 |
| Kimi K3 | low | 53.6 | $2.37 | $1.58 | 6 |
| GLM 5.3 Flash | max | 52.4 | $0.58 | $0.097 | 103 |
One task is an average task of Artificial Analysis’ Intelligence Index, at each effort it was measured at. Through the plan it is the API cost scaled by what the plan charges against its estimated allowance, so it holds only if you use the whole allowance. Effort changes how fast an allowance runs out, not its size. No allowance estimate exists for GLM 5.2, DeepSeek V4.1 Flash, Qwen3.7 Plus, MiMo-V2.5, Hy3, MiMo-V2.5-Pro, Kimi K2.7 Code, LongCat 2.0 and Qwen3.8 Flash on this plan.
What a real agent turn costs
A turn is one prompt run to the end, every tool call and re-read of the context included, averaged over T3 Code’s users and priced at API list price. The last column divides the allowance's API value by that price.
| Model | Tokens per turn | From cache | API price per turn | Turns a month on this plan |
|---|---|---|---|---|
| Claude Haiku 5.5 | 1.7 million | 96% | $0.030 | 494 |
| GPT-6 Luna | 3.0 million | 96% | $0.046 | 329 |
Who it suits
- Trying OpenCode on a small budget: it is the cheapest paid tier that includes the coding tool.
- Switching between many models: it lists 23.
Every OpenCode tier
| Tier | Price | Allowance at full use | Per dollar |
|---|---|---|---|
| OpenCode Go | $10 | About $170 on DeepSeek V4.1 Flash (estimate) | 17× |
| OpenCode Go Teams | $20 | No figure | – |
| OpenCode Go Plus | $40 | No figure | – |
Against the closest plans
The nearest plan in price from each of three other vendors.
| Plan | Price | Allowance at full use | Per dollar | Five-hour limit | Best model |
|---|---|---|---|---|---|
| OpenCode Go | $10 | About $170 on DeepSeek V4.1 Flash (estimate) | 17× | Not stated | Kimi K3.1 Pro |
| GitHub Copilot Pro | $10 | $15 of credits, published | 1.5× | Not stated | GPT-6.1 Sol |
| $8 | No figure | – | Not stated | GPT-6 Luna | |
| $8 | No figure | – | Not stated | Grok 4.6 |
Head to head
What changed
- 30 Oct 2026 — Added Go Teams at $20 per member. Source
OpenCode Go: questions people ask
- How much does OpenCode Go cost?
- OpenCode Go costs $10 a month. Price read from OpenCode's pricing page on 30 Oct 2026.
- What are OpenCode Go's usage limits?
- OpenCode describes the allowance as “Curated, affordable models; per-model requests per 5 hours and monthly usage published”. How it refreshes: 5-hour and monthly limits.
- Which models does OpenCode Go include?
- Kimi K3.1 Pro, Kimi K3, GLM 5.2, GLM 5.3, Qwen3.8 Max, DeepSeek V4.1 Flash, GPT-5.6 Luna, GLM 5.3 Flash, Qwen3.7 Plus, MiMo-V2.6-Pro, Grok 4.7, Claude Haiku 5.5, MiMo-V2.5, MiniMax M3, DeepSeek V4 Pro, Hy3, MiMo-V2.5-Pro, GPT-6 Luna, MiMo-V2.6-Flash, MiniMax M3.1 Flash, Kimi K2.7 Code, LongCat 2.0 and Qwen3.8 Flash. The highest-ranked is Kimi K3.1 Pro, at 61.8 on the BenchLeader Index.
- Is OpenCode Go worth it?
- Used in full on DeepSeek V4.1 Flash, the allowance comes to about 10 billion tokens a month by Real API Pricing's medium-confidence estimate: about $170 at API prices, 17× the subscription. Against paying per token, it pays for itself after about 42 agent sessions a month on Kimi K3. Every figure assumes the whole allowance is used; an unused month makes every task dearer.
- OpenCode Go or GitHub Copilot Pro?
- Both cost $10 a month. OpenCode Go goes further per dollar: at full use it is worth about 17× its price in API usage on DeepSeek V4.1 Flash, against 1.5× for GitHub Copilot Pro in credits.
Sources
- Price and the allowance wording: opencode.ai, read 30 Oct 2026.
- Limits: OpenCode Go docs, read 30 Oct 2026.
- Allowance estimate: Real API Pricing (MIT licence), data to 30 Oct 2026, priced at BenchLeader’s API prices of 16 Nov 2026.
- Cost per index task at each effort: Artificial Analysis, as of 16 Nov 2026.
- Real agent turns: T3 Code’s usage telemetry as published by Slopalytics, priced at API list price.
Every plan side by side, with how the estimates are made, is on the coding plans page.
Cite as: BenchLeader, “OpenCode Go: limits, models and real cost”, data as of 16 Nov 2026, https://www.benchleader.com/coding-plans/opencode-go.