OpenCode
Open-source, model-agnostic terminal coding agent supporting 75+ providers.
OpenCode is an open-source, model-agnostic terminal coding agent (from Anomaly, ex-SST) supporting 75+ providers. It's free to run as an OSS tool with your own API key.
Vendor details
Canonical URL
https://opencode.ai
Category
Coding agent
Subcategory
Coding — CLI agent (OSS)
Use cases & customers
Primary use cases
Target customers
Deployment options
In practice
You want a coding agent in your terminal that isn't tied to one vendor's model. OpenCode is model-agnostic, supporting more than 75 providers with your own API key.
Closed coding tools mean trusting a black box with your code. OpenCode is open source, so you can run and inspect it yourself.
You'd rather pay providers directly than a markup. OpenCode is free to run as an open-source tool, with cost only on the model you choose.
Agentic Index coverage score
7.5 / 14 capabilities · 54%
| Integrations & Tool CallingAgent Features research report + JSON Feature Rubric | Partial |
|---|---|
| Workflow OrchestrationAgent Features research report + JSON Feature Rubric | Partial |
| Knowledge Grounding & RAGAgent Features research report + JSON Feature Rubric | Partial |
| Human Oversight & GuardrailsAgent Features research report + JSON Feature Rubric | Full |
| Security, Identity & GovernanceAgent Features research report + JSON Feature Rubric | Partial |
| Observability & AuditabilityAgent Features research report + JSON Feature Rubric | Partial |
| Memory & State PersistenceAgent Features research report + JSON Feature Rubric | Partial |
| Deployment & Data ResidencyAgent Features research report + JSON Feature Rubric | Full |
| Prebuilt Agents, Templates & PacksAgent Features research report + JSON Feature Rubric | Unable to verify |
| Triggers & Channel CoverageAgent Features research report + JSON Feature Rubric | Unable to verify |
| Model Flexibility & RoutingAgent Features research report + JSON Feature Rubric | Full |
| APIs, SDKs & MCP ExtensibilityAgent Features research report + JSON Feature Rubric | Full |
| Testing, Debugging & OptimizationAgent Features research report + JSON Feature Rubric | Partial |
| Browser & Computer UseAgent Features research report + JSON Feature Rubric | Unable to verify |
The Agentic Index coverage score grades every vendor Full, Partial or Unable to verify against the same 14 buyer facing capabilities, from public evidence only. Each capability links to how all vendors in the index score on it. How this evidence is graded
Recent platform changes
OpenCode version 1.18.2 introduces default execution constraints that prevent subagents from independently launching nested subagents. Administrators and developers can override this safeguard by configuring a specific subagent_depth limit when deeper recursive delegation is necessary.
Bears on: Workflow orchestration
View sourceOpenCode released version 1.18.0, finalizing the Desktop v2 migration. The update introduces a redesigned review panel with persistent file tabs, a new composer menu for adding context without losing draft text, and per-prompt model selection. A transition setting is included to allow users to temporarily toggle between the legacy and new interfaces.
Bears on: Workflow orchestration
View sourceOpenCode 1.17.11 added session snapshots and rollback, letting users revert a session to an earlier message and undo file changes. The release also brought file-based agent loading, skill discovery, and expanded OpenAI model support via AWS Bedrock (building on the MCP resource tooling and minimal CLI mode shipped the day before in 1.17.10).
Bears on: Memory / state
View sourcePricing
Free / OSS
hybrid
Included quota
Free (BYOK): full OpenCode features, any provider/local model — no included quota, you fund tokens. Go ($10/mo): curated open-weight models with a monthly dollar-value usage allowance (~$60 of usage / ~17K requests on lighter models).
What is public
The core OpenCode agent (by SST/Anomaly) is free and open source (MIT), BYO key across 75+ providers and local models — you pay only your own model tokens. Optional managed add-ons launched in 2026: OpenCode Zen (pay-as-you-go gateway, prepaid balance e.g. $20, per-request billing), OpenCode Go ($5 first month then $10/mo for curated open-weight models), and OpenCode Black (enterprise gateway, reported from ~$20/mo).
Billing mechanics
The tool itself is free; cost comes from model access. Three optional paths: (1) BYO API key — pay your provider directly; (2) Zen — a curated model gateway billed pay-as-you-go per request against a prepaid balance (auto-reloads $20 when below $5; set workspace/per-member monthly limits); (3) Go — a flat $10/mo subscription bundling curated open-weight models with dollar-value usage limits, falling back to Zen balance when exhausted.
Cost watchouts
BYO-key model token spend, Zen auto-reload (can exceed your set monthly limit), Go limited to curated (mostly open-weight/Chinese-lab) models, Go falls back to a separate $20 Zen top-up, enterprise (Black) pricing unannounced
Variable cost rationale
Free core has zero subscription floor but uncapped BYO-key token spend; Zen is pure PAYG; only Go's flat $10 caps cost (until fallback) — so exposure is usage-driven unless you stay on Go
Additional watchouts
The free core is genuinely free (BYO key); the paid Zen/Go/Black layers are optional and newer — Go is curated open-weight models only, no Western proprietary models
Overage / add-ons
Go usage is capped by dollar value; on exhaustion it blocks or falls back to a Zen pay-as-you-go balance. Zen auto-reloads $20 when the balance drops below $5 (configurable; can be disabled). BYO-key usage bills directly from your provider with no cap.
Sales call required
No, self serve available
Free / trial
Free (OSS self-host)
Commercial notes
MIT-licensed, ~160K+ GitHub stars, ~7.5M monthly devs; the paid tiers launched after Anthropic's Jan 2026 block on third-party Claude-subscription auth; Go targets international users (US/EU/Singapore hosting); zero-retention on managed models
Key ambiguities
Go's exact request limits depend on model choice (dollar-denominated); Zen per-token rates and Black/enterprise tiers aren't fully published
Cancellation / refund
Cancel any time; Zen/Go let you set monthly spend limits; prepaid balances; no long-term contract
Support SLA / resale
Community/OSS support on the free core; managed support via Zen/Go/Black (enterprise gateway, reported volume discounts)
Missing data
OpenCode Zen per-token rates and OpenCode Black (enterprise) tier pricing aren't fully published; Go limits are dollar-denominated and vary by model
Related vendors
- Cognition — Maker of Devin, an autonomous AI software engineer
- 10Web — Agentic website platform whose team of ten specialized AI agents…
- Ageiro — Early stage enterprise platform that turns business goals into…
- Aider — Open-source, model-agnostic terminal coding agent that works…
- Anthropic Claude Code — Anthropic's terminal and IDE-native coding agent with sub-agent…
- AppFactor — Agentic orchestration platform that autonomously maintains and…
Alternatives to OpenCode
The closest documented capability profiles to OpenCode among coding agents tracked by Agentic Index, ordered by similarity on the same 14 point evidence the rankings use. No vendor pays for placement.
- Aider6.5 / 14A lighter documented profile than OpenCodeOpenCode vs Aider →
- Charm10.0 / 14Adds documented Prebuilt Agents, Templates & PacksOpenCode vs Charm →
- Goose9.0 / 14Adds documented Prebuilt Agents, Templates & Packs and Triggers & Channel Coverage
- iGent7.0 / 14Fuller documented coverage on Workflow Orchestration and Testing, Debugging & Optimization
- Warp10.0 / 14Adds documented Prebuilt Agents, Templates & Packs and Triggers & Channel Coverage
- Cline10.5 / 14Adds documented Prebuilt Agents, Templates & Packs and Triggers & Channel Coverage
Similarity is computed from each vendor's Agentic Index coverage score evidence, axis by axis, not from the totals. How this evidence is graded