Skip to content

opencode.md

Latest commit

 

History

History
104 lines (69 loc) · 4.07 KB

opencode.md

File metadata and controls

104 lines (69 loc) · 4.07 KB
Raw

OpenCode — Free LLM API Config

Last verified: October 2026. Free-tier models change often — check the provider pages on free-llm.com before relying on a specific model.

Point OpenCode at any free OpenAI-compatible backend in 30 seconds.

How It Works

OpenCode supports multiple LLM providers through environment variables or the /connect command. When you configure a free provider, OpenCode routes API calls through that backend instead of paid options.

Quick Config

Groq (fast inference, no credit card)

export GROQ_API_KEY="gsk_your_groq_api_key"

Get your key at console.groq.com/keys. Recommended model: openai/gpt-oss-120b (llama-3.3-70b-versatile was retired from Groq's free tier on August 16, 2026).

Google AI Studio (free tier, generous limits)

export GOOGLE_API_KEY="your_google_ai_studio_key"

Get your key at aistudio.google.com. Recommended model: gemini-3.8-flash (Gemini 2.0 was shut down on June 1, 2026; Pro models are no longer on the free tier).

OpenRouter (~14 free models, single key, no credit card)

export OPENROUTER_API_KEY="sk-or-v1-your_openrouter_key"

Get your key at openrouter.ai/keys. Recommended model: nvidia/nemotron-3-ultra-550b-a55b:free (the free catalog rotates quickly — see the official free models list). Some free models may use your prompts for training.

Mistral (La Plateforme)

export MISTRAL_API_KEY="your_mistral_key"

Get your key at console.mistral.ai/api-keys. Recommended model: codestral-latest for coding tasks (open-mistral-nemo was retired on July 31, 2026).

Cerebras (very fast inference, $5 trial credit)

export CEREBRAS_API_KEY="your_cerebras_key"

Get your key at cerebras.ai/inference. Recommended model: gpt-oss-120b (llama-3.3-70b is deprecated). Since August 2026 a verified payment method is required to activate the $5 trial credit (valid 30 days).

Using /connect Command

OpenCode has a built-in /connect command that simplifies provider setup:

  1. Run /connect in the OpenCode TUI
  2. Select your provider (opencode, groq, google, etc.)
  3. Follow the prompts to enter your API key
  4. Start coding with free models!

Persistent Config

Add to your shell profile (~/.zshrc or ~/.bashrc):

# Free LLM API backend for OpenCode
export GROQ_API_KEY="gsk_your_key_here"
# OR
export GOOGLE_API_KEY="your_key_here"
# OR
export OPENROUTER_API_KEY="sk-or-v1-your_key_here"

OpenCode vs Other Coding Assistants

Feature OpenCode Claude Code Cursor Codex CLI
Open Source ✅ ❌ ❌ ❌
Free Models ✅ (via providers) ❌ ❌ ❌
Terminal + Desktop + IDE ✅ Terminal only Desktop only Terminal only
GitHub Copilot Support ✅ ❌ ❌ ❌
ChatGPT Plus/Pro Support ✅ ❌ ❌ ❌
Multi-session ✅ ❌ ❌ ❌
Share Links ✅ ❌ ❌ ❌

Recommended Free Setup

For the best free experience with OpenCode:

  1. Get a Groq API key (fastest, no credit card required)
  2. Install OpenCode: curl -fsSL https://opencode.ai/install | bash
  3. Configure: Run /connect, select groq, paste your key
  4. Start coding: OpenCode will use GPT OSS 120B for free

Caveats

  • Free tiers have rate limits. Groq: ~1,000 requests/day on gpt-oss-120b. OpenRouter: 50 requests/day (1,000 with $10 of lifetime credits).
  • OpenCode works best with models that support tool use well (GPT OSS 120B, Qwen3.8 27B, Nemotron 3 Ultra).
  • The /connect command makes setup easier than manual environment variables.

More Providers

See the full Provider Directory and Quick Reference in the main README for all 55+ providers, or browse free-llm.com.