You're paying premium prices for every request—whether it's a complex coding task or a simple "acknowledged." Multi-model routing sends each request to the right model for its complexity level, cutting costs dramatically.
The Problem: One Model for Everything
Most OpenClaw users configure a single model for all tasks. This means:
- Simple tasks (heartbeats, acknowledgments, status checks) → Premium model pricing
- Medium tasks (email summaries, simple Q&A) → Premium model pricing
- Complex tasks (coding, debugging, multi-step analysis) → Premium model pricing ✓
Only the complex tasks actually benefit from a premium model. Everything else is overpaying.
The Solution: Tiered Routing (Coming Soon)
Note: This feature is defined in ClawBridge's roadmap and will be available in a future release. This article explains the concept and how it will work.
ClawBridge's Diagnostic A10 will analyze your historical token usage and recommend a multi-model routing configuration:
4-Tier Routing Plan
| Tier | Request Type | Suggested Model | Cost Level |
|---|---|---|---|
| Tier 1 (Free) | Heartbeats, simple checks | Local Ollama (7B) | $0.00 |
| Tier 2 (Budget) | Simple responses, acknowledgments | Claude Haiku / GPT-4o-mini | Lowest |
| Tier 3 (Standard) | Daily tasks, emails, summaries | Claude Sonnet / GPT-4o | Mid |
| Tier 4 (Premium) | Coding, reasoning, orchestration | Claude Opus / GPT-4 | Highest |
How It Will Work
- Analysis: ClawBridge reviews your past 7 days of usage and categorizes each request by complexity.
- Distribution map: Shows what percentage of your requests fall into each tier.
- Routing recommendation: Suggests specific models for each tier based on your provider configuration.
- Savings projection: Calculates exact dollar savings based on your actual usage patterns.
Estimated Impact
For a typical user with 500 requests/day:
| Distribution | Requests/Day | Before (All Sonnet) | After (Tiered) |
|---|---|---|---|
| Tier 1 (30%) | 150 | $0.68 | $0.00 (local) |
| Tier 2 (30%) | 150 | $0.68 | $0.08 (Haiku) |
| Tier 3 (30%) | 150 | $0.68 | $0.68 (Sonnet) |
| Tier 4 (10%) | 50 | $0.23 | $0.75 (Opus) |
| Total | 500 | $2.27/day | $1.51/day |
| Monthly | $68.10 | $45.30 |
Savings: ~$22.80/month (33%). With local routing for Tier 1, savings increase to 50–70%.
Trade-offs
- Routing accuracy: Incorrect classification could send a complex task to a cheap model, producing lower quality output. The routing logic must be well-tuned.
- Fallback chain: If a lower-tier model fails or produces poor results, the system should automatically escalate to a higher tier.
- Configuration complexity: Multi-model setups have more moving parts. More models = more provider credentials to manage.
- Provider fallback ordering: Direct-API providers (Anthropic, Google API key) should be prioritized before proxy providers (OpenRouter, KiloCode) to avoid single points of failure.
Prepare Now
While this feature is in development:
- Use A01 to switch to a cheaper default model — the simplest version of this concept.
- Use A02 to reduce heartbeat frequency — fewer requests means less to route.
- Use A08 to set up local Ollama for the free tier when ready.
ClawBridge is free and open source (MIT License) — install it in seconds, own it forever. Get ClawBridge Free →