XClawRouter
Smart LLM router that saves 84% on inference costs by routing each request to the cheapest model that can handle it. 78 models across 9 providers (6 free models), all through one wallet.
Install
openclaw plugins install @blockrun/xclawrouter
Setup
# Enable smart routing (auto-picks cheapest model per request)
openclaw models set blockrun/auto
# Or pin a specific model
openclaw models set openai/gpt-4o
How Routing Works
XClawRouter classifies each request into one of four tiers:
- SIMPLE — factual lookups, greetings, translations → gemini-2.5-flash ($0.30/$2.50)
- MEDIUM — summaries, explanations, data extraction → kimi-k3 ($3/$15)
- COMPLEX — code generation, multi-step analysis → gemini-3.1-pro ($2/$12)
- REASONING — proofs, formal logic, multi-step math → deepseek-reasoner ($0.14/$0.28)
Prices are per 1M input/output tokens on the default auto profile. Per-tier
savings percentages are deliberately not quoted: the published figure is blended
across a stated workload mix, and a per-tier number invites comparison against a
baseline nobody wrote down.
Rules handle 80% of requests in <1ms. Only ambiguous queries hit the LLM classifier ($0.00003 per classification).
Available Models
Example Output
[XClawRouter] google/gemini-2.5-flash (SIMPLE, rules, confidence=0.92)
Cost: $0.0025 | Baseline: $0.308 | Saved: 99.2%
1---2name: xclawrouter3description: Smart LLM router for OKX — agentic wallet, 66 models, x402 micropayments on Base + Solana. Routes every request to the cheapest capable model. 11 free NVIDIA models included.4---56# XClawRouter78Smart LLM router that saves <!-- br:savings.autoVsBaselinePct -->84<!-- /br:savings.autoVsBaselinePct -->% on inference costs by routing each request to the cheapest model that can handle it. <!-- br:models.chatVisible -->78<!-- /br:models.chatVisible --> models across 9 providers (<!-- br:models.free -->6<!-- /br:models.free --> free models), all through one wallet.910## Install1112```bash13openclaw plugins install @blockrun/xclawrouter14```1516## Setup1718```bash19# Enable smart routing (auto-picks cheapest model per request)20openclaw models set blockrun/auto2122# Or pin a specific model23openclaw models set openai/gpt-4o24```2526## How Routing Works2728XClawRouter classifies each request into one of four tiers:2930- **SIMPLE** — factual lookups, greetings, translations → gemini-2.5-flash ($0.30/$2.50)31- **MEDIUM** — summaries, explanations, data extraction → kimi-k3 ($3/$15)32- **COMPLEX** — code generation, multi-step analysis → gemini-3.1-pro ($2/$12)33- **REASONING** — proofs, formal logic, multi-step math → deepseek-reasoner ($0.14/$0.28)3435Prices are per 1M input/output tokens on the default `auto` profile. Per-tier36savings percentages are deliberately not quoted: the published figure is blended37across a stated workload mix, and a per-tier number invites comparison against a38baseline nobody wrote down.3940Rules handle ~~80% of requests in <1ms. Only ambiguous queries hit the LLM classifier (~~$0.00003 per classification).4142## Available Models4344<!-- br:models.chatVisible -->78<!-- /br:models.chatVisible --> models including: claude-fable-5, claude-opus-5, claude-sonnet-5, gpt-5.6-terra, gpt-5.6-sol, gpt-5.5, gpt-5.4, gemini-3.1-pro, gemini-3.6-flash, grok-4.5, grok-4.3, glm-5.3, kimi-k3, qwen3.7-max, mimo-v2.5, deepseek-v4-pro, deepseek-chat, and the free models (nemotron-3.5-lightning [default], nemotron-3-nano-30b, laguna-xs-2.1, north-mini-code, nemotron-3-nano-omni-30b-a3b-reasoning [vision], nemotron-3-ultra-550b, llama-3.2-11b-vision [vision]).4546## Example Output4748```49[XClawRouter] google/gemini-2.5-flash (SIMPLE, rules, confidence=0.92)50 Cost: $0.0025 | Baseline: $0.308 | Saved: 99.2%51```
BlockRunAI