{
  "schemaVersion": "1.0",
  "item": {
    "slug": "openclaw-model-router",
    "name": "Model Router",
    "source": "tencent",
    "type": "skill",
    "category": "AI 智能",
    "sourceUrl": "https://clawhub.ai/jsgoecke/openclaw-model-router",
    "canonicalUrl": "https://clawhub.ai/jsgoecke/openclaw-model-router",
    "targetPlatform": "OpenClaw"
  },
  "install": {
    "downloadMode": "redirect",
    "downloadUrl": "/downloads/openclaw-model-router",
    "sourceDownloadUrl": "https://wry-manatee-359.convex.site/api/v1/download?slug=openclaw-model-router",
    "sourcePlatform": "tencent",
    "targetPlatform": "OpenClaw",
    "installMethod": "Manual import",
    "extraction": "Extract archive",
    "prerequisites": [
      "OpenClaw"
    ],
    "packageFormat": "ZIP package",
    "includedAssets": [
      "SKILL.md"
    ],
    "primaryDoc": "SKILL.md",
    "quickSetup": [
      "Download the package from Yavira.",
      "Extract the archive and review SKILL.md first.",
      "Import or place the package into your OpenClaw setup."
    ],
    "agentAssist": {
      "summary": "Hand the extracted package to your coding agent with a concrete install brief instead of figuring it out manually.",
      "steps": [
        "Download the package from Yavira.",
        "Extract it into a folder your agent can access.",
        "Paste one of the prompts below and point your agent at the extracted folder."
      ],
      "prompts": [
        {
          "label": "New install",
          "body": "I downloaded a skill package from Yavira. Read SKILL.md from the extracted folder and install it by following the included instructions. Tell me what you changed and call out any manual steps you could not complete."
        },
        {
          "label": "Upgrade existing",
          "body": "I downloaded an updated skill package from Yavira. Read SKILL.md from the extracted folder, compare it with my current installation, and upgrade it while preserving any custom configuration unless the package docs explicitly say otherwise. Summarize what changed and any follow-up checks I should run."
        }
      ]
    },
    "sourceHealth": {
      "source": "tencent",
      "status": "healthy",
      "reason": "direct_download_ok",
      "recommendedAction": "download",
      "checkedAt": "2026-04-30T16:55:25.780Z",
      "expiresAt": "2026-05-07T16:55:25.780Z",
      "httpStatus": 200,
      "finalUrl": "https://wry-manatee-359.convex.site/api/v1/download?slug=network",
      "contentType": "application/zip",
      "probeMethod": "head",
      "details": {
        "probeUrl": "https://wry-manatee-359.convex.site/api/v1/download?slug=network",
        "contentDisposition": "attachment; filename=\"network-1.0.0.zip\"",
        "redirectLocation": null,
        "bodySnippet": null
      },
      "scope": "source",
      "summary": "Source download looks usable.",
      "detail": "Yavira can redirect you to the upstream package for this source.",
      "primaryActionLabel": "Download for OpenClaw",
      "primaryActionHref": "/downloads/openclaw-model-router"
    },
    "validation": {
      "installChecklist": [
        "Use the Yavira download entry.",
        "Review SKILL.md after the package is downloaded.",
        "Confirm the extracted package contains the expected setup assets."
      ],
      "postInstallChecks": [
        "Confirm the extracted package includes the expected docs or setup files.",
        "Validate the skill or prompts are available in your target agent workspace.",
        "Capture any manual follow-up steps the agent could not complete."
      ]
    },
    "downloadPageUrl": "https://openagent3.xyz/downloads/openclaw-model-router",
    "agentPageUrl": "https://openagent3.xyz/skills/openclaw-model-router/agent",
    "manifestUrl": "https://openagent3.xyz/skills/openclaw-model-router/agent.json",
    "briefUrl": "https://openagent3.xyz/skills/openclaw-model-router/agent.md"
  },
  "agentAssist": {
    "summary": "Hand the extracted package to your coding agent with a concrete install brief instead of figuring it out manually.",
    "steps": [
      "Download the package from Yavira.",
      "Extract it into a folder your agent can access.",
      "Paste one of the prompts below and point your agent at the extracted folder."
    ],
    "prompts": [
      {
        "label": "New install",
        "body": "I downloaded a skill package from Yavira. Read SKILL.md from the extracted folder and install it by following the included instructions. Tell me what you changed and call out any manual steps you could not complete."
      },
      {
        "label": "Upgrade existing",
        "body": "I downloaded an updated skill package from Yavira. Read SKILL.md from the extracted folder, compare it with my current installation, and upgrade it while preserving any custom configuration unless the package docs explicitly say otherwise. Summarize what changed and any follow-up checks I should run."
      }
    ]
  },
  "documentation": {
    "source": "clawhub",
    "primaryDoc": "SKILL.md",
    "sections": [
      {
        "title": "Smart Model Router",
        "body": "Intelligent cost-aware model routing for OpenClaw agents.\n\nBefore executing any task via sessions_spawn or delegating to a sub-agent, classify the task complexity using the rules below and route to the optimal model. This saves 60-90% on LLM costs by using cheap models for simple work and reserving premium models for tasks that genuinely need them."
      },
      {
        "title": "Core Principle",
        "body": "Route every request to the cheapest model that can handle it well."
      },
      {
        "title": "Step 1: Classify Task Complexity",
        "body": "Score the task on these dimensions. Count how many COMPLEX/REASONING indicators are present:"
      },
      {
        "title": "SIMPLE indicators (route to Tier 1)",
        "body": "Greetings, small talk, status checks, heartbeats\nSingle factual questions (\"What is X?\", \"Define Y\")\nSimple translations, format conversions\nFile lookups, directory listings, basic shell commands\nCalendar checks, weather queries\nTasks under 50 tokens with no technical depth\nKeywords: \"what is\", \"define\", \"translate\", \"list\", \"check\", \"hello\", \"status\""
      },
      {
        "title": "MODERATE indicators (route to Tier 2)",
        "body": "Summarization of documents or conversations\nSingle-file code edits, bug fixes, simple refactors\nWriting emails, messages, short-form content\nData extraction, parsing, formatting\nExplaining concepts, answering \"how to\" questions\nResearch requiring synthesis of a few sources\nKeywords: \"summarize\", \"explain\", \"write\", \"fix this\", \"how to\", \"extract\""
      },
      {
        "title": "COMPLEX indicators (route to Tier 3)",
        "body": "Multi-file code generation or refactoring\nArchitecture design, system design\nCreative writing (stories, long-form, nuanced tone)\nDebugging complex issues across multiple systems\nAnalysis requiring multiple perspectives\nTasks with constraints (\"optimize for X while maintaining Y\")\nKeywords: \"build\", \"design\", \"architect\", \"refactor\", \"create\", \"implement\", \"analyze\""
      },
      {
        "title": "REASONING indicators (route to Tier 4)",
        "body": "Mathematical proofs, formal logic\nMulti-step reasoning chains (\"first X, then Y, therefore Z\")\nSecurity vulnerability analysis\nPerformance optimization with tradeoffs\nScientific analysis, hypothesis testing\nAny task with 2+ of: \"prove\", \"derive\", \"why does\", \"compare and contrast\", \"evaluate tradeoffs\", \"step by step\"\nKeywords: \"prove\", \"derive\", \"reason\", \"why does\", \"evaluate\", \"theorem\""
      },
      {
        "title": "Special Rules",
        "body": "2+ reasoning keywords → always Tier 4 (high confidence)\nCode blocks or multi-file references → minimum Tier 2\n\"Debug\" + stack traces → Tier 3\nHeartbeats and /status → always Tier 1\nWhen uncertain, default to Tier 2 (fast, cheap, good enough)"
      },
      {
        "title": "Tier 0 — FREE (OpenRouter free tier)",
        "body": "ModelCostBest ForGemini 2.5 Flash (free)$0.00High-volume simple tasks, translationGemini 2.5 Flash-Lite (free)$0.00Translation, marketingGemini 3 Flash Preview (free)$0.00Technology, health, scienceDeepSeek V3.2 (free)$0.00Roleplay, creative writingMoonshot Kimi K2.5 (free)$0.00Technology, programmingArcee Trinity Large Preview (free)$0.00Creative writing, storytelling, agents\n\nDefault Tier 0 model: openrouter/free (auto-selects from available free models)\n\nAccess via OpenRouter with model IDs like google/gemini-2.5-flash, deepseek/deepseek-v3.2-20251201, moonshotai/kimi-k2.5-0127. Or use openrouter/free to auto-route across all free models.\n\nNote: Free models have rate limits and may have variable availability. Use for non-critical tasks only."
      },
      {
        "title": "Tier 1 — SIMPLE (near-zero cost)",
        "body": "ModelInput $/MTokOutput $/MTokBest ForGemini 2.0 Flash$0.10$0.40Default simple tier — fast, multimodal, 1M contextGPT-4o-mini$0.15$0.60Simple tasks, multimodalGPT-5 Nano$0.05$0.40Cheapest OpenAI optionDeepSeek V3$0.27$1.10Budget general-purposeGemini 2.5 Flash-Lite$0.10$0.40Most economical Google model\n\nDefault Tier 1 model: gemini-2.0-flash (best cost/reliability balance)"
      },
      {
        "title": "Tier 2 — MODERATE (balanced)",
        "body": "ModelInput $/MTokOutput $/MTokBest ForClaude Haiku 4.5$1.00$5.00Near-frontier, fast, great codingGPT-4o$2.50$10.00Multimodal, tool use, solid all-rounderGemini 2.5 Flash$0.15$0.60Thinking-enabled, fast reasoningGPT-5 Mini$0.25$2.00Balanced performance, 400K contextMistral Medium 3$0.40$2.00European languages, balanced\n\nDefault Tier 2 model: claude-haiku-4-5 (best quality-to-price at this tier)"
      },
      {
        "title": "Tier 3 — COMPLEX (premium)",
        "body": "ModelInput $/MTokOutput $/MTokBest ForClaude Sonnet 4.5$3.00$15.00Best coding-to-cost ratio, most popularGPT-5$1.25$10.00Flagship coding and agentic tasksGPT-5.3 Codex$1.75*$14.00*Most capable agentic coding modelGemini 2.5 Pro$1.25$10.00Coding, reasoning, up to 2M contextClaude Opus 4.5$5.00$25.00Maximum intelligence, agentic tasksGrok 4$3.00$15.00Frontier reasoning, real-time data\n\n*GPT-5.3 Codex API pricing not yet officially released; estimated from GPT-5.2 Codex rates.\n\nDefault Tier 3 model: claude-sonnet-4-5 (best balance of quality, coding, and cost)"
      },
      {
        "title": "Tier 4 — REASONING (maximum capability)",
        "body": "ModelInput $/MTokOutput $/MTokBest ForClaude Opus 4.6$5.00$25.00Latest frontier reasoning, extended thinking, 1M context (beta)Claude Opus 4.5$5.00$25.00Extended thinking, frontier reasoningo3$2.00$8.00Deep STEM reasoningDeepSeek R1$0.55$2.19Budget reasoning (20-50x cheaper than o1)o4-mini$1.10$4.40Efficient reasoning\n\nDefault Tier 4 model: claude-opus-4-6 with extended thinking enabled"
      },
      {
        "title": "🟢 Balanced Mode (DEFAULT)",
        "body": "Use the default model for each tier as listed above. Escalate to next tier if the model produces low-quality output or fails."
      },
      {
        "title": "🔵 Aggressive Mode (Maximum Savings)",
        "body": "Override tier defaults to cheapest option:\n\nTier 0-1: openrouter/free ($0.00) for simple tasks, fall back to gemini-2.0-flash ($0.10/$0.40)\nTier 2: gemini-2.5-flash ($0.15/$0.60)\nTier 3: gemini-2.5-pro ($1.25/$10.00)\nTier 4: deepseek-r1 ($0.55/$2.19)\n\nSavings: 70-99% vs always using Opus"
      },
      {
        "title": "🟡 Quality Mode (Maximum Quality)",
        "body": "Override tier defaults to best-in-class:\n\nTier 1: claude-haiku-4-5 ($1.00/$5.00)\nTier 2: claude-sonnet-4-5 ($3.00/$15.00)\nTier 3: claude-opus-4-6 ($5.00/$25.00) or gpt-5.3-codex for coding\nTier 4: claude-opus-4-6 ($5.00/$25.00) with extended thinking"
      },
      {
        "title": "Step 4: Execute with sessions_spawn",
        "body": "# Simple task — Tier 1\nsessions_spawn --task \"What's on my calendar today?\" --model gemini-2.0-flash\n\n# Moderate task — Tier 2\nsessions_spawn --task \"Summarize this document\" --model claude-haiku-4-5\n\n# Complex task — Tier 3\nsessions_spawn --task \"Build a React auth component with tests\" --model claude-sonnet-4-5\n\n# Reasoning task — Tier 4\nsessions_spawn --task \"Prove this algorithm is O(n log n)\" --model claude-opus-4-6"
      },
      {
        "title": "Progressive Escalation Pattern",
        "body": "When uncertain about complexity, start cheap and escalate:\n\n# 1. Try Tier 1 with timeout\nsessions_spawn --task \"Fix this bug\" --model gemini-2.0-flash --runTimeoutSeconds 60\n\n# 2. If output is poor or times out, escalate to Tier 2\nsessions_spawn --task \"Fix this bug\" --model claude-haiku-4-5\n\n# 3. If still failing, escalate to Tier 3\nsessions_spawn --task \"Fix this complex bug\" --model claude-sonnet-4-5\n\nMaximum escalation chain: 3 attempts. If Tier 3 fails, surface the error to the user rather than burning tokens."
      },
      {
        "title": "Parallel Processing for Batch Tasks",
        "body": "Route batch/parallel tasks to Tier 1 models for massive savings:\n\n# Batch summaries in parallel with cheap model\nsessions_spawn --task \"Summarize doc A\" --model gemini-2.0-flash &\nsessions_spawn --task \"Summarize doc B\" --model gemini-2.0-flash &\nsessions_spawn --task \"Summarize doc C\" --model gemini-2.0-flash &\nwait\n\n# Then analyze results with premium model\nsessions_spawn --task \"Synthesize findings from all summaries\" --model claude-sonnet-4-5"
      },
      {
        "title": "Special Routing Rules",
        "body": "ScenarioRoute ToWhyHeartbeat / status checkTier 0 (openrouter/free) or Tier 1Zero intelligence needed, save every centVision / image analysisgemini-2.5-proBest multimodal + huge contextLong context (>100K tokens)gemini-2.5-pro or gpt-51M-2M context windowsChinese language tasksdeepseek-v3 or glm-4.7Optimized for ChineseReal-time web data neededgrok-4.1-fastBuilt-in X/web search, 2M contextAgentic coding tasksgpt-5.3-codex or claude-sonnet-4-5Purpose-built for agentic code workflowsCode generationclaude-sonnet-4-5 minimumBest code quality per dollarMath / formal proofso3 or claude-opus-4-6 with thinkingSpecialized reasoning"
      },
      {
        "title": "Cost Comparison (Typical Workload)",
        "body": "For a typical OpenClaw day (24 heartbeats + 20 sub-agent tasks + 10 user queries):\n\nStrategyMonthly CostSavingsAll Opus 4.6~$200baselineSmart routing (balanced)~$4578%Smart routing (aggressive)~$1592%Smart routing (aggressive + free tier)~$597%All free models (OpenRouter)~$0100% (but rate-limited & unreliable)"
      },
      {
        "title": "When NOT to Route Down",
        "body": "Always use Tier 3+ for:\n\nSecurity-sensitive code review\nFinancial calculations where errors are costly\nArchitecture decisions that affect the whole codebase\nAnything the user explicitly asks for premium quality\nTasks where the user says \"be thorough\" or \"take your time\""
      },
      {
        "title": "Mode Switching",
        "body": "Users can switch modes mid-conversation:\n\n\"Use aggressive routing\" → Switch to cheapest models per tier\n\"Use quality mode\" → Switch to best models per tier\n\"Use balanced routing\" → Return to defaults\n\"Use [specific model] for this\" → Override routing for one task"
      },
      {
        "title": "Pricing Reference (February 2026)",
        "body": "All prices per million tokens. Models are listed from cheapest to most expensive output:\n\nModelInputOutputContextProviderOpenRouter Free Models$0.00$0.00VariesOpenRouterGPT-5 Nano$0.05$0.40400KOpenAIGemini 2.0 Flash$0.10$0.401MGoogleGemini 2.5 Flash-Lite$0.10$0.401MGoogleGPT-4o-mini$0.15$0.60128KOpenAIGemini 2.5 Flash$0.15$0.601MGoogleGrok 4.1 Fast$0.20$0.502MxAIGPT-5 Mini$0.25$2.00400KOpenAIDeepSeek V3$0.27$1.1064KDeepSeekDeepSeek R1$0.55$2.1964KDeepSeekClaude Haiku 4.5$1.00$5.00200KAnthropico4-mini$1.10$4.40200KOpenAIGemini 2.5 Pro$1.25$10.001MGoogleGPT-5$1.25$10.00400KOpenAIGPT-5.3 Codex$1.75*$14.00*400KOpenAIo3$2.00$8.00200KOpenAIGPT-4o$2.50$10.00128KOpenAIClaude Sonnet 4.5$3.00$15.00200KAnthropicGrok 4$3.00$15.00256KxAIClaude Opus 4.5$5.00$25.00200KAnthropicClaude Opus 4.6$5.00$25.00200K (1M beta)Anthropic\n\n*GPT-5.3 Codex pricing estimated from GPT-5.2 Codex; official API pricing pending.\n\nNote: Prices change. Check provider pricing pages for current rates. Batch API discounts (50% off) and prompt caching (50-90% off) can reduce costs further. OpenRouter free models have rate limits — see openrouter.ai/collections/free-models for current availability."
      }
    ],
    "body": "Smart Model Router\n\nIntelligent cost-aware model routing for OpenClaw agents.\n\nBefore executing any task via sessions_spawn or delegating to a sub-agent, classify the task complexity using the rules below and route to the optimal model. This saves 60-90% on LLM costs by using cheap models for simple work and reserving premium models for tasks that genuinely need them.\n\nCore Principle\n\nRoute every request to the cheapest model that can handle it well.\n\nStep 1: Classify Task Complexity\n\nScore the task on these dimensions. Count how many COMPLEX/REASONING indicators are present:\n\nSIMPLE indicators (route to Tier 1)\nGreetings, small talk, status checks, heartbeats\nSingle factual questions (\"What is X?\", \"Define Y\")\nSimple translations, format conversions\nFile lookups, directory listings, basic shell commands\nCalendar checks, weather queries\nTasks under 50 tokens with no technical depth\nKeywords: \"what is\", \"define\", \"translate\", \"list\", \"check\", \"hello\", \"status\"\nMODERATE indicators (route to Tier 2)\nSummarization of documents or conversations\nSingle-file code edits, bug fixes, simple refactors\nWriting emails, messages, short-form content\nData extraction, parsing, formatting\nExplaining concepts, answering \"how to\" questions\nResearch requiring synthesis of a few sources\nKeywords: \"summarize\", \"explain\", \"write\", \"fix this\", \"how to\", \"extract\"\nCOMPLEX indicators (route to Tier 3)\nMulti-file code generation or refactoring\nArchitecture design, system design\nCreative writing (stories, long-form, nuanced tone)\nDebugging complex issues across multiple systems\nAnalysis requiring multiple perspectives\nTasks with constraints (\"optimize for X while maintaining Y\")\nKeywords: \"build\", \"design\", \"architect\", \"refactor\", \"create\", \"implement\", \"analyze\"\nREASONING indicators (route to Tier 4)\nMathematical proofs, formal logic\nMulti-step reasoning chains (\"first X, then Y, therefore Z\")\nSecurity vulnerability analysis\nPerformance optimization with tradeoffs\nScientific analysis, hypothesis testing\nAny task with 2+ of: \"prove\", \"derive\", \"why does\", \"compare and contrast\", \"evaluate tradeoffs\", \"step by step\"\nKeywords: \"prove\", \"derive\", \"reason\", \"why does\", \"evaluate\", \"theorem\"\nSpecial Rules\n2+ reasoning keywords → always Tier 4 (high confidence)\nCode blocks or multi-file references → minimum Tier 2\n\"Debug\" + stack traces → Tier 3\nHeartbeats and /status → always Tier 1\nWhen uncertain, default to Tier 2 (fast, cheap, good enough)\nStep 2: Select Model from Tier\nTier 0 — FREE (OpenRouter free tier)\nModel\tCost\tBest For\nGemini 2.5 Flash (free)\t$0.00\tHigh-volume simple tasks, translation\nGemini 2.5 Flash-Lite (free)\t$0.00\tTranslation, marketing\nGemini 3 Flash Preview (free)\t$0.00\tTechnology, health, science\nDeepSeek V3.2 (free)\t$0.00\tRoleplay, creative writing\nMoonshot Kimi K2.5 (free)\t$0.00\tTechnology, programming\nArcee Trinity Large Preview (free)\t$0.00\tCreative writing, storytelling, agents\n\nDefault Tier 0 model: openrouter/free (auto-selects from available free models)\n\nAccess via OpenRouter with model IDs like google/gemini-2.5-flash, deepseek/deepseek-v3.2-20251201, moonshotai/kimi-k2.5-0127. Or use openrouter/free to auto-route across all free models.\n\nNote: Free models have rate limits and may have variable availability. Use for non-critical tasks only.\n\nTier 1 — SIMPLE (near-zero cost)\nModel\tInput $/MTok\tOutput $/MTok\tBest For\nGemini 2.0 Flash\t$0.10\t$0.40\tDefault simple tier — fast, multimodal, 1M context\nGPT-4o-mini\t$0.15\t$0.60\tSimple tasks, multimodal\nGPT-5 Nano\t$0.05\t$0.40\tCheapest OpenAI option\nDeepSeek V3\t$0.27\t$1.10\tBudget general-purpose\nGemini 2.5 Flash-Lite\t$0.10\t$0.40\tMost economical Google model\n\nDefault Tier 1 model: gemini-2.0-flash (best cost/reliability balance)\n\nTier 2 — MODERATE (balanced)\nModel\tInput $/MTok\tOutput $/MTok\tBest For\nClaude Haiku 4.5\t$1.00\t$5.00\tNear-frontier, fast, great coding\nGPT-4o\t$2.50\t$10.00\tMultimodal, tool use, solid all-rounder\nGemini 2.5 Flash\t$0.15\t$0.60\tThinking-enabled, fast reasoning\nGPT-5 Mini\t$0.25\t$2.00\tBalanced performance, 400K context\nMistral Medium 3\t$0.40\t$2.00\tEuropean languages, balanced\n\nDefault Tier 2 model: claude-haiku-4-5 (best quality-to-price at this tier)\n\nTier 3 — COMPLEX (premium)\nModel\tInput $/MTok\tOutput $/MTok\tBest For\nClaude Sonnet 4.5\t$3.00\t$15.00\tBest coding-to-cost ratio, most popular\nGPT-5\t$1.25\t$10.00\tFlagship coding and agentic tasks\nGPT-5.3 Codex\t$1.75*\t$14.00*\tMost capable agentic coding model\nGemini 2.5 Pro\t$1.25\t$10.00\tCoding, reasoning, up to 2M context\nClaude Opus 4.5\t$5.00\t$25.00\tMaximum intelligence, agentic tasks\nGrok 4\t$3.00\t$15.00\tFrontier reasoning, real-time data\n\n*GPT-5.3 Codex API pricing not yet officially released; estimated from GPT-5.2 Codex rates.\n\nDefault Tier 3 model: claude-sonnet-4-5 (best balance of quality, coding, and cost)\n\nTier 4 — REASONING (maximum capability)\nModel\tInput $/MTok\tOutput $/MTok\tBest For\nClaude Opus 4.6\t$5.00\t$25.00\tLatest frontier reasoning, extended thinking, 1M context (beta)\nClaude Opus 4.5\t$5.00\t$25.00\tExtended thinking, frontier reasoning\no3\t$2.00\t$8.00\tDeep STEM reasoning\nDeepSeek R1\t$0.55\t$2.19\tBudget reasoning (20-50x cheaper than o1)\no4-mini\t$1.10\t$4.40\tEfficient reasoning\n\nDefault Tier 4 model: claude-opus-4-6 with extended thinking enabled\n\nStep 3: Apply Optimization Mode\n🟢 Balanced Mode (DEFAULT)\n\nUse the default model for each tier as listed above. Escalate to next tier if the model produces low-quality output or fails.\n\n🔵 Aggressive Mode (Maximum Savings)\n\nOverride tier defaults to cheapest option:\n\nTier 0-1: openrouter/free ($0.00) for simple tasks, fall back to gemini-2.0-flash ($0.10/$0.40)\nTier 2: gemini-2.5-flash ($0.15/$0.60)\nTier 3: gemini-2.5-pro ($1.25/$10.00)\nTier 4: deepseek-r1 ($0.55/$2.19)\n\nSavings: 70-99% vs always using Opus\n\n🟡 Quality Mode (Maximum Quality)\n\nOverride tier defaults to best-in-class:\n\nTier 1: claude-haiku-4-5 ($1.00/$5.00)\nTier 2: claude-sonnet-4-5 ($3.00/$15.00)\nTier 3: claude-opus-4-6 ($5.00/$25.00) or gpt-5.3-codex for coding\nTier 4: claude-opus-4-6 ($5.00/$25.00) with extended thinking\nStep 4: Execute with sessions_spawn\n# Simple task — Tier 1\nsessions_spawn --task \"What's on my calendar today?\" --model gemini-2.0-flash\n\n# Moderate task — Tier 2\nsessions_spawn --task \"Summarize this document\" --model claude-haiku-4-5\n\n# Complex task — Tier 3\nsessions_spawn --task \"Build a React auth component with tests\" --model claude-sonnet-4-5\n\n# Reasoning task — Tier 4\nsessions_spawn --task \"Prove this algorithm is O(n log n)\" --model claude-opus-4-6\n\nProgressive Escalation Pattern\n\nWhen uncertain about complexity, start cheap and escalate:\n\n# 1. Try Tier 1 with timeout\nsessions_spawn --task \"Fix this bug\" --model gemini-2.0-flash --runTimeoutSeconds 60\n\n# 2. If output is poor or times out, escalate to Tier 2\nsessions_spawn --task \"Fix this bug\" --model claude-haiku-4-5\n\n# 3. If still failing, escalate to Tier 3\nsessions_spawn --task \"Fix this complex bug\" --model claude-sonnet-4-5\n\n\nMaximum escalation chain: 3 attempts. If Tier 3 fails, surface the error to the user rather than burning tokens.\n\nParallel Processing for Batch Tasks\n\nRoute batch/parallel tasks to Tier 1 models for massive savings:\n\n# Batch summaries in parallel with cheap model\nsessions_spawn --task \"Summarize doc A\" --model gemini-2.0-flash &\nsessions_spawn --task \"Summarize doc B\" --model gemini-2.0-flash &\nsessions_spawn --task \"Summarize doc C\" --model gemini-2.0-flash &\nwait\n\n# Then analyze results with premium model\nsessions_spawn --task \"Synthesize findings from all summaries\" --model claude-sonnet-4-5\n\nSpecial Routing Rules\nScenario\tRoute To\tWhy\nHeartbeat / status check\tTier 0 (openrouter/free) or Tier 1\tZero intelligence needed, save every cent\nVision / image analysis\tgemini-2.5-pro\tBest multimodal + huge context\nLong context (>100K tokens)\tgemini-2.5-pro or gpt-5\t1M-2M context windows\nChinese language tasks\tdeepseek-v3 or glm-4.7\tOptimized for Chinese\nReal-time web data needed\tgrok-4.1-fast\tBuilt-in X/web search, 2M context\nAgentic coding tasks\tgpt-5.3-codex or claude-sonnet-4-5\tPurpose-built for agentic code workflows\nCode generation\tclaude-sonnet-4-5 minimum\tBest code quality per dollar\nMath / formal proofs\to3 or claude-opus-4-6 with thinking\tSpecialized reasoning\nCost Comparison (Typical Workload)\n\nFor a typical OpenClaw day (24 heartbeats + 20 sub-agent tasks + 10 user queries):\n\nStrategy\tMonthly Cost\tSavings\nAll Opus 4.6\t~$200\tbaseline\nSmart routing (balanced)\t~$45\t78%\nSmart routing (aggressive)\t~$15\t92%\nSmart routing (aggressive + free tier)\t~$5\t97%\nAll free models (OpenRouter)\t~$0\t100% (but rate-limited & unreliable)\nWhen NOT to Route Down\n\nAlways use Tier 3+ for:\n\nSecurity-sensitive code review\nFinancial calculations where errors are costly\nArchitecture decisions that affect the whole codebase\nAnything the user explicitly asks for premium quality\nTasks where the user says \"be thorough\" or \"take your time\"\nMode Switching\n\nUsers can switch modes mid-conversation:\n\n\"Use aggressive routing\" → Switch to cheapest models per tier\n\"Use quality mode\" → Switch to best models per tier\n\"Use balanced routing\" → Return to defaults\n\"Use [specific model] for this\" → Override routing for one task\nPricing Reference (February 2026)\n\nAll prices per million tokens. Models are listed from cheapest to most expensive output:\n\nModel\tInput\tOutput\tContext\tProvider\nOpenRouter Free Models\t$0.00\t$0.00\tVaries\tOpenRouter\nGPT-5 Nano\t$0.05\t$0.40\t400K\tOpenAI\nGemini 2.0 Flash\t$0.10\t$0.40\t1M\tGoogle\nGemini 2.5 Flash-Lite\t$0.10\t$0.40\t1M\tGoogle\nGPT-4o-mini\t$0.15\t$0.60\t128K\tOpenAI\nGemini 2.5 Flash\t$0.15\t$0.60\t1M\tGoogle\nGrok 4.1 Fast\t$0.20\t$0.50\t2M\txAI\nGPT-5 Mini\t$0.25\t$2.00\t400K\tOpenAI\nDeepSeek V3\t$0.27\t$1.10\t64K\tDeepSeek\nDeepSeek R1\t$0.55\t$2.19\t64K\tDeepSeek\nClaude Haiku 4.5\t$1.00\t$5.00\t200K\tAnthropic\no4-mini\t$1.10\t$4.40\t200K\tOpenAI\nGemini 2.5 Pro\t$1.25\t$10.00\t1M\tGoogle\nGPT-5\t$1.25\t$10.00\t400K\tOpenAI\nGPT-5.3 Codex\t$1.75*\t$14.00*\t400K\tOpenAI\no3\t$2.00\t$8.00\t200K\tOpenAI\nGPT-4o\t$2.50\t$10.00\t128K\tOpenAI\nClaude Sonnet 4.5\t$3.00\t$15.00\t200K\tAnthropic\nGrok 4\t$3.00\t$15.00\t256K\txAI\nClaude Opus 4.5\t$5.00\t$25.00\t200K\tAnthropic\nClaude Opus 4.6\t$5.00\t$25.00\t200K (1M beta)\tAnthropic\n\n*GPT-5.3 Codex pricing estimated from GPT-5.2 Codex; official API pricing pending.\n\nNote: Prices change. Check provider pricing pages for current rates. Batch API discounts (50% off) and prompt caching (50-90% off) can reduce costs further. OpenRouter free models have rate limits — see openrouter.ai/collections/free-models for current availability."
  },
  "trust": {
    "sourceLabel": "tencent",
    "provenanceUrl": "https://clawhub.ai/jsgoecke/openclaw-model-router",
    "publisherUrl": "https://clawhub.ai/jsgoecke/openclaw-model-router",
    "owner": "jsgoecke",
    "version": "0.1.0",
    "license": null,
    "verificationStatus": "Indexed source record"
  },
  "links": {
    "detailUrl": "https://openagent3.xyz/skills/openclaw-model-router",
    "downloadUrl": "https://openagent3.xyz/downloads/openclaw-model-router",
    "agentUrl": "https://openagent3.xyz/skills/openclaw-model-router/agent",
    "manifestUrl": "https://openagent3.xyz/skills/openclaw-model-router/agent.json",
    "briefUrl": "https://openagent3.xyz/skills/openclaw-model-router/agent.md"
  }
}