# Send AI Intelligence Hub - Real-time Model Capability Tracking to your agent Hand the extracted package to your coding agent with a concrete install brief instead of figuring it out manually. ## Fast path - 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. ## Suggested prompts ### New install ```text I downloaded a skill package from Yavira. Read SKILL.md from the extracted folder and install it by following the included instructions. Then review README.md for any prerequisites, environment setup, or post-install checks. Tell me what you changed and call out any manual steps you could not complete. ``` ### Upgrade existing ```text 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. Then review README.md for any prerequisites, environment setup, or post-install checks. Summarize what changed and any follow-up checks I should run. ``` ## Machine-readable fields ```json { "schemaVersion": "1.0", "item": { "slug": "model-benchmarks", "name": "AI Intelligence Hub - Real-time Model Capability Tracking", "source": "tencent", "type": "skill", "category": "ๆ•ˆ็Ž‡ๆๅ‡", "sourceUrl": "https://clawhub.ai/Notestone/model-benchmarks", "canonicalUrl": "https://clawhub.ai/Notestone/model-benchmarks", "targetPlatform": "OpenClaw" }, "install": { "downloadUrl": "/downloads/model-benchmarks", "sourceDownloadUrl": "https://wry-manatee-359.convex.site/api/v1/download?slug=model-benchmarks", "sourcePlatform": "tencent", "targetPlatform": "OpenClaw", "packageFormat": "ZIP package", "primaryDoc": "SKILL.md", "includedAssets": [ "README.md", "SKILL.md", "scripts/run.py", "benchmarks/latest.json", "benchmarks/2026-03-01.json", "examples/daily-optimization.sh" ], "downloadMode": "redirect", "sourceHealth": { "source": "tencent", "slug": "model-benchmarks", "status": "healthy", "reason": "direct_download_ok", "recommendedAction": "download", "checkedAt": "2026-04-28T23:48:50.655Z", "expiresAt": "2026-05-05T23:48:50.655Z", "httpStatus": 200, "finalUrl": "https://wry-manatee-359.convex.site/api/v1/download?slug=model-benchmarks", "contentType": "application/zip", "probeMethod": "head", "details": { "probeUrl": "https://wry-manatee-359.convex.site/api/v1/download?slug=model-benchmarks", "contentDisposition": "attachment; filename=\"model-benchmarks-1.0.0.zip\"", "redirectLocation": null, "bodySnippet": null, "slug": "model-benchmarks" }, "scope": "item", "summary": "Item download looks usable.", "detail": "Yavira can redirect you to the upstream package for this item.", "primaryActionLabel": "Download for OpenClaw", "primaryActionHref": "/downloads/model-benchmarks" }, "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." ] } }, "links": { "detailUrl": "https://openagent3.xyz/skills/model-benchmarks", "downloadUrl": "https://openagent3.xyz/downloads/model-benchmarks", "agentUrl": "https://openagent3.xyz/skills/model-benchmarks/agent", "manifestUrl": "https://openagent3.xyz/skills/model-benchmarks/agent.json", "briefUrl": "https://openagent3.xyz/skills/model-benchmarks/agent.md" } } ``` ## Documentation ### ๐Ÿง  Model Benchmarks - Global AI Intelligence Hub "Know thy models, optimize thy costs" โ€” Real-time AI capability tracking for intelligent compute routing ### ๐ŸŽฏ What It Does Transform your OpenClaw deployment from guessing to data-driven model selection: ๐Ÿ” Real-time Intelligence โ€” Pulls latest capability data from LMSYS Arena, BigCode, HuggingFace leaderboards ๐Ÿ“Š Standardized Scoring โ€” Unified 0-100 capability scores across coding, reasoning, creative tasks ๐Ÿ’ฐ Cost Efficiency โ€” Calculates performance-per-dollar ratios to find hidden gems ๐ŸŽฏ Smart Recommendations โ€” Suggests optimal models for specific task types ๐Ÿ“ˆ Trend Analysis โ€” Tracks model performance changes over time ### ๐Ÿš€ Why You Need This Problem: OpenClaw users often overpay for AI by using expensive models for simple tasks, or underperform by using cheap models for complex work. Solution: This skill provides real-time model intelligence to route tasks optimally: ็ฟป่ฏ‘ไปปๅŠก: Gemini 2.0 Flash (445x cost efficiency vs Claude) ๅคๆ‚็ผ–็จ‹: Claude 3.5 Sonnet (92/100 coding score) ็ฎ€ๅ•้—ฎ็ญ”: GPT-4o Mini (85x cheaper than GPT-4) Result: Users report 60-95% cost reduction with maintained or improved quality. ### Install & First Run # Fetch latest model intelligence python3 skills/model-benchmarks/scripts/run.py fetch # Find best model for your task python3 skills/model-benchmarks/scripts/run.py recommend --task coding # Check any model's capabilities python3 skills/model-benchmarks/scripts/run.py query --model gpt-4o ### Sample Output ๐Ÿ† Top 3 recommendations for coding: 1. gemini-2.0-flash Task Score: 81.5/100 Cost Efficiency: 445.33 Avg Price: $0.19/1M tokens 2. claude-3.5-sonnet Task Score: 92.0/100 Cost Efficiency: 10.28 Avg Price: $9.00/1M tokens ### With OpenClaw Model Routing # Get optimal model, then configure OpenClaw BEST_MODEL=$(python3 skills/model-benchmarks/scripts/run.py recommend --task coding --json | jq -r '.models[0]') openclaw config set agents.defaults.model.primary "$BEST_MODEL" ### Daily Intelligence Updates # Add to crontab for fresh data 0 8 * * * cd ~/.openclaw/workspace && python3 skills/model-benchmarks/scripts/run.py fetch ### Cost Monitoring Dashboard # Generate cost efficiency report python3 skills/model-benchmarks/scripts/run.py analyze --export-csv > model_costs.csv ### ๐Ÿ“Š Supported Data Sources PlatformCoverageUpdate FrequencyCapabilities TrackedLMSYS Chatbot Arena100+ modelsDailyGeneral, Reasoning, CreativeBigCode Leaderboard50+ modelsWeeklyCoding (HumanEval, MBPP)Open LLM Leaderboard200+ modelsDailyKnowledge, ComprehensionAlpaca Eval80+ modelsWeeklyInstruction Following ### ๐ŸŽฏ Task-to-Model Mapping The skill intelligently maps your tasks to optimal models: Task TypePrimary CapabilityRecommended ModelscodingCoding + ReasoningGemini 2.0 Flash, Claude 3.5 SonnetwritingCreative + GeneralClaude 3.5 Sonnet, GPT-4oanalysisReasoning + ComprehensionGPT-4o, Claude 3.5 SonnettranslationGeneral + KnowledgeGemini 2.0 Flash, GPT-4o MinimathReasoning + KnowledgeGPT-4o, Claude 3.5 SonnetsimpleGeneralGemini 2.0 Flash, GPT-4o Mini ### Cost Optimization Workflow Profile your tasks โ€” What do you do most often? Get recommendations โ€” Run analysis for each task type Configure routing โ€” Set up model fallbacks Monitor & adjust โ€” Weekly intelligence updates ### Finding Hidden Gems # Discover undervalued models python3 skills/model-benchmarks/scripts/run.py analyze --sort-by efficiency --limit 10 ### Trend Analysis # Compare model performance over time python3 skills/model-benchmarks/scripts/run.py trends --model gpt-4o --days 30 ### Custom Benchmark Sources Edit BENCHMARK_SOURCES in scripts/run.py to add new evaluation platforms. ### Task-Specific Scoring Customize TASK_CAPABILITY_MAP to weight capabilities for your specific use cases. ### Enterprise Integration Slack alerts for model price changes API endpoints for programmatic access Custom dashboards with exported JSON data ### ๐Ÿ“ˆ Real-World Results Startups using this skill report: ๐Ÿ—๏ธ Dev Teams: 78% cost reduction by routing simple tasks to Gemini 2.0 Flash ๐Ÿ“ Content Agencies: 65% savings using task-specific model routing ๐Ÿ”ฌ Research Labs: 45% efficiency gain with capability-driven model selection ### ๐Ÿ›ก๏ธ Privacy & Security No personal data collected โ€” Only public benchmark results Local processing โ€” All analysis runs on your machine Optional caching โ€” Benchmark data cached locally for faster queries No external dependencies โ€” Uses only Python standard library ### ๐Ÿ”ฎ Roadmap v1.1: Real-time price monitoring from OpenRouter/Anthropic APIs v1.2: Custom benchmark suite for your specific tasks v1.3: Multi-provider cost comparison (OpenRouter vs Direct APIs) v2.0: Predictive model performance based on task characteristics ### ๐Ÿค Contributing Found a new benchmark platform? Want to improve the scoring algorithm? Fork the skill on GitHub Add your enhancement Submit a pull request Help the OpenClaw community optimize their AI costs! ### ๐Ÿ“ž Support Documentation: Full API reference in scripts/run.py --help Issues: Report bugs or request features via GitHub Community: Join discussions on OpenClaw Discord Examples: More integration examples in examples/ directory Make every token count โ€” choose your models wisely! ๐Ÿง  ## Trust - Source: tencent - Verification: Indexed source record - Publisher: Notestone - Version: 1.0.0 ## Source health - Status: healthy - Item download looks usable. - Yavira can redirect you to the upstream package for this item. - Health scope: item - Reason: direct_download_ok - Checked at: 2026-04-28T23:48:50.655Z - Expires at: 2026-05-05T23:48:50.655Z - Recommended action: Download for OpenClaw ## Links - [Detail page](https://openagent3.xyz/skills/model-benchmarks) - [Send to Agent page](https://openagent3.xyz/skills/model-benchmarks/agent) - [JSON manifest](https://openagent3.xyz/skills/model-benchmarks/agent.json) - [Markdown brief](https://openagent3.xyz/skills/model-benchmarks/agent.md) - [Download page](https://openagent3.xyz/downloads/model-benchmarks)