Feature: Smart Model Fallback System

[KingLabsA]

Feature Request: Smart Model Fallback System

Summary

Implement automatic model fallback when rate limits are hit, with intelligent model selection based on capability requirements.

Problem Statement

Users with multiple LLM providers (Anthropic, OpenAI, Google, DeepSeek, LM Studio) experience workflow interruptions when one provider hits rate limits (429 errors). Currently, manual intervention is required to switch providers.

Proposed Solution

Core Feature: Auto Model Fallback

{
  "model": {
    "primary": "anthropic/claude-sonnet-4-20250514",
    "fallback": [
      "openai/gpt-5",
      "google/gemini-2-flash",
      "deepseek/deepseek-coder"
    ],
    "onRateLimit": "auto-switch",
    "retryAttempts": 3
  }
}

Smart Model Selection

Models have varying capabilities. The system should match requirements:

Capability	Description
tools	Function calling / tool use
image	Vision / image analysis
reasoning	Chain-of-thought / reasoning
thinking	Extended thinking (like o1, Claude extended)
uncensored	Full capability (no content filtering)
context	Max context window
speed	Response latency
cost	Cost tier

Use Cases

1. Developer with rate limits: Auto-switch from Claude to GPT-5
2. Image analysis needed: Fall back to model with vision
3. Heavy reasoning task: Switch to model with thinking capability
4. Budget optimization: Use cheap models first, expensive only when needed
5. Local models: Use LM Studio when cloud is unavailable

Suggested Implementation

1. Built-in Config

Add model.fallback array to config schema

2. Capability Database

Maintain model capability registry (similar to models.dev)

3. Auto-detection

• Detect 429, 503, rate limit errors
• Query fallback chain in priority order
• Match capability requirements

4. User Tools

• smart_model_select(requirements) - programmatic selection
• /models - show all with capabilities
• Status bar indicator of current provider

Benefits

• Zero-interruption workflow
• Better reliability for production
• Cost optimization
• Differentiates from competitors

Priority

Medium-High

Alternative: Plugin Approach (Current)

A reference implementation exists as an OpenCode plugin:

• Detection of rate limit errors
• Smart fallback based on capability matching
• Tools for programmatic model selection

References

• OpenCode Config Docs
• Models.dev - Model capability database
• Related: Variant system for model configuration

---

Feature request by: Community

[github-actions]

This issue doesn't fully meet our contributing guidelines.

What needs to be fixed:

• The title must start with [FEATURE]: as required by the feature request template (current title: "Feature: Smart Model Fallback System")
• The issue reads as an AI-generated wall of text — a large structured design document with implementation tables, JSON config samples, use case lists, and capability registries is not acceptable. Please replace this with a concise, human-written description of the problem you're experiencing and what you'd like to see.
• "Feature request by: Community" appears to be placeholder/generic attribution and should be removed or replaced with real context about your use case.

Please edit this issue to address the above within 2 hours, or it will be automatically closed.

---

This issue might be a duplicate of existing issues. Please check:

• [FEATURE]:Add model fallback/failover support #25150: Add model fallback/failover support
• [FEATURE]: Native Model Fallback / Failover Support #7602: Native Model Fallback / Failover Support
• [FEATURE]: Configurable provider/model fallback on transient errors #20100: Configurable provider/model fallback on transient errors
• [FEATURE]: Allow agent-level model pools and fallback policies #16868: Allow agent-level model pools and fallback policies

If you believe this was flagged incorrectly, please let a maintainer know.

[github-actions]

This issue has been automatically closed because it was not updated to meet our contributing guidelines within the 2-hour window.

Feel free to open a new issue that follows our issue templates.