# Send Models 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. 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. Summarize what changed and any follow-up checks I should run. ``` ## Machine-readable fields ```json { "schemaVersion": "1.0", "item": { "slug": "models", "name": "Models", "source": "tencent", "type": "skill", "category": "AI 智能", "sourceUrl": "https://clawhub.ai/ivangdavila/models", "canonicalUrl": "https://clawhub.ai/ivangdavila/models", "targetPlatform": "OpenClaw" }, "install": { "downloadUrl": "/downloads/models", "sourceDownloadUrl": "https://wry-manatee-359.convex.site/api/v1/download?slug=models", "sourcePlatform": "tencent", "targetPlatform": "OpenClaw", "packageFormat": "ZIP package", "primaryDoc": "SKILL.md", "includedAssets": [ "SKILL.md" ], "downloadMode": "redirect", "sourceHealth": { "source": "tencent", "slug": "models", "status": "healthy", "reason": "direct_download_ok", "recommendedAction": "download", "checkedAt": "2026-05-02T11:59:41.944Z", "expiresAt": "2026-05-09T11:59:41.944Z", "httpStatus": 200, "finalUrl": "https://wry-manatee-359.convex.site/api/v1/download?slug=models", "contentType": "application/zip", "probeMethod": "head", "details": { "probeUrl": "https://wry-manatee-359.convex.site/api/v1/download?slug=models", "contentDisposition": "attachment; filename=\"models-1.0.0.zip\"", "redirectLocation": null, "bodySnippet": null, "slug": "models" }, "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/models" }, "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/models", "downloadUrl": "https://openagent3.xyz/downloads/models", "agentUrl": "https://openagent3.xyz/skills/models/agent", "manifestUrl": "https://openagent3.xyz/skills/models/agent.json", "briefUrl": "https://openagent3.xyz/skills/models/agent.md" } } ``` ## Documentation ### Core Principle No single model is best for everything — match model to task, not brand loyalty A $0.75/M model often performs identically to a $40/M model for simple tasks Test cheaper alternatives before committing to expensive defaults ### Cost Reality Output tokens cost 3-10x more than input tokens — advertised input prices are misleading Calculate real cost with your actual input/output ratio, not theoretical pricing Batch/async APIs offer 50% discounts — use them for non-real-time workloads Prompt caching reduces repeated context costs significantly ### Coding Architecture and design decisions: Use frontier models (Opus-class) — they catch subtle issues cheaper models miss Day-to-day implementation: Mid-tier models (Sonnet-class) offer 90% of capability at 20% of cost Parallel subtasks and scaffolding: Fast/cheap models (Haiku-class) — speed matters more than depth Code review: Thorough models catch async bugs and edge cases that fast models miss ### Non-Coding Complex reasoning and math: Extended thinking modes justify their cost for hard problems General assistance: User preference studies favor models different from benchmark leaders High-volume simple queries: Cheapest models perform identically — don't overpay Long documents: Context window size determines viability — some offer 1M+ tokens ### Claude Code vs Codex CLI Claude Code: Fast iteration, UI/frontend, interactive debugging — developer stays in the loop Codex CLI: Long-running background tasks, large refactors, set-and-forget — accuracy over speed Both tools have value — use Claude Code for implementation, Codex for final review File size limits differ — Claude Code struggles with files over 25K tokens ### Orchestration Pattern Planning phase: Use expensive/smart models to break down problems correctly Execution phase: Use balanced models, parallelize where possible Review phase: Use accurate models for final verification — catches bugs others miss This pattern beats using one model for everything at similar total cost ### Benchmark Skepticism Benchmark scores vary 2-3x based on scaffolding and evaluation method User preference rankings differ significantly from benchmark rankings SWE-bench scores don't predict real-world coding quality reliably Models drift week-to-week — last month's best may underperform today ### Open Source Viability DeepSeek and similar models approach frontier performance at 1/50th API cost Self-hosting eliminates API rate limits and price variability MIT/Apache licensed models allow commercial use without restrictions Consider for: data privacy, cost predictability, custom fine-tuning ### Model Selection Mistakes Using premium models for chatbot responses that cheap models handle identically Ignoring context window limits — chunking long documents costs more than using large-context models Expecting consistency — same prompt gives different results over time as models update Trusting speed over accuracy for complex tasks — fast models trade thoroughness for latency ### Practical Guidelines Default to mid-tier for most tasks, escalate to frontier only when quality suffers Track actual costs per workflow, not just per-token rates Build verification into pipelines — don't trust any model blindly Reassess model choices quarterly — pricing and capabilities shift constantly ## Trust - Source: tencent - Verification: Indexed source record - Publisher: ivangdavila - 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: 2026-05-02T11:59:41.944Z - Expires at: 2026-05-09T11:59:41.944Z - Recommended action: Download for OpenClaw ## Links - [Detail page](https://openagent3.xyz/skills/models) - [Send to Agent page](https://openagent3.xyz/skills/models/agent) - [JSON manifest](https://openagent3.xyz/skills/models/agent.json) - [Markdown brief](https://openagent3.xyz/skills/models/agent.md) - [Download page](https://openagent3.xyz/downloads/models)