[VickyXAI]

Stale model counts that scripts/sync-brand-numbers.mjs cannot reach, checked against https://blockrun.ai/brand/numbers.json (chatVisible 82, totalVisible 109, free 6).

File	Old → new	What it claims	Form
skills/clawrouter/SKILL.md (body, "models including…")	7 free open-weight models → 6	BlockRun free models (models.free)	marker br:models.free
same line	removed nemotron-3-nano-30b [fastest, ~121 tok/s] from the named list	the gateway delisted nvidia/nemotron-3-nano-30b on 2026-09-08; the six remaining names are exactly the live $0 set in /api/v1/models	text
src/index.ts header comment	55+ models → 82 models	BlockRun chat catalog (chatVisible)	literal
src/index.ts plugin description	55+ models → 82 models	BlockRun chat catalog	literal
src/index.ts registration log	(55+ models via x402) → (${VISIBLE_OPENCLAW_MODELS.length} models via x402)	what the provider actually registers (the picker subset, currently 56); counted from code, so it cannot drift	computed
src/models.ts header comment	Maps BlockRun's 55+ AI models → Maps BlockRun's model catalog	the file's own BLOCKRUN_MODELS (114 entries incl. deprecated); a number here would just go stale again	number removed
openclaw.plugin.json description	55+ models → 82 models	BlockRun chat catalog	literal (JSON, no markers)
docs/clawrouter-cuts-llm-api-costs-500x.md img alt and code fence	marker → plain 82 (value unchanged)	BlockRun chat catalog	literal: a marker inside an HTML attribute or a code fence isn't a comment, so it shows up as visible text

Not changed:

• dist/** still says 55+; it is build output and regenerates on the next release build.
• src/proxy.ts FREE_MODELS still lists free/nemotron-3-nano-30b. loadGatewayCatalog() already skips rungs the gateway doesn't serve, so it's harmless at runtime, but it's a dead rung that someone should remove in a follow-up.
• The "78% cost savings" (index.ts) and "92% cost savings" (openclaw.plugin.json) figures disagree with savings.autoVsBaselinePct (84). This PR only covers model counts.
• Dated posts in docs/ (benchmark: 39/46 models; the 2026-03-20 classifier post: 55+; the 2026-02-03 plan: 30+) are snapshots and stay as they are.

Checks: node scripts/sync-brand-numbers.mjs --check passes, tsc --noEmit passes, eslint passes on the touched files, prettier --check passes, and the full vitest run passes (1109 passed, 1 skipped), including src/router/brand-numbers.test.ts.

🤖 Generated with Claude Code

Summary by CodeRabbit

• Updates
  • Updated the advertised model count to 82.
  • Revised the free-model list, replacing one model with a reasoning model that supports a 256K context window.
  • Clarified that free-tier access is text-only; image requests use a paid vision model.

[coderabbitai]

Navigate logical layers of code changes, visualize relationships, and explore their blast radius.

Note

Currently processing new changes in this PR. This may take a few minutes, please wait...

⚙️ Run configuration

Configuration used: Repository: BlockRunAI/ClawRouter/.coderabbit.yaml

Review profile: CHILL

Plan: Advanced

Run ID: 25aca6ee-d763-4f21-9c74-b245ea35b9dd

📥 Commits

Reviewing files that changed from the base of the PR and between 1feca68 and 7bb169a.

📒 Files selected for processing (4)

• openclaw.plugin.json
• skills/clawrouter/SKILL.md
• src/index.ts
• src/models.ts

 ______________________________________________________
< Please wait while I pretend to understand your code. >
 ------------------------------------------------------
  \
   \   (\__/)
       (•ㅅ•)
       / 　 づ

✨ Finishing Touches

📝 Generate docstrings

• Commit to this branch
• Create a new PR

🧪 Generate unit tests (beta)

• Commit to this branch
• Create a new PR

---

Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out.

❤️ Share

• X
• Mastodon
• Reddit
• LinkedIn

_{Comment @coderabbitai help to get the list of available commands.}