@blockrun/clawrouter 0.12.258 → 0.12.260

package/README.md

@@ -34,7 +34,7 @@ Agents can only sign transactions.<br><br>
 
 </div>
 
-> **ClawRouter** is an open-source smart LLM router that reduces AI API costs by up to <!-- br:savings.autoVsBaselinePct -->88<!-- /br:savings.autoVsBaselinePct -->%. It analyzes each request across <!-- br:clawrouter.dimensions -->15<!-- /br:clawrouter.dimensions --> dimensions and routes to the cheapest capable model in under 1ms, entirely locally. ClawRouter is the only LLM router built for autonomous AI agents — it uses wallet signatures for authentication (no API keys) and USDC micropayments via the x402 protocol (no credit cards). <!-- br:models.chatVisible -->73<!-- /br:models.chatVisible --> models from OpenAI, Anthropic, Google, xAI, DeepSeek, and more. MIT licensed.
+> **ClawRouter** is an open-source smart LLM router that reduces AI API costs by up to <!-- br:savings.autoVsBaselinePct -->84<!-- /br:savings.autoVsBaselinePct -->%. It analyzes each request across <!-- br:clawrouter.dimensions -->15<!-- /br:clawrouter.dimensions --> dimensions and routes to the cheapest capable model in under 1ms, entirely locally. ClawRouter is the only LLM router built for autonomous AI agents — it uses wallet signatures for authentication (no API keys) and USDC micropayments via the x402 protocol (no credit cards). <!-- br:models.chatVisible -->76<!-- /br:models.chatVisible --> models from OpenAI, Anthropic, Google, xAI, DeepSeek, and more. MIT licensed.
 
 ---
 
@@ -46,7 +46,7 @@ Every other LLM router was built for **human developers** — create an account,
 
 ClawRouter is built for the agent-first world:
 
-- **Starts at $0** — <!-- br:models.free -->7<!-- /br:models.free --> open-weight models are free forever (incl. a 1M-context reasoner, the vision-capable Nemotron Omni, and two sub-second coders)
+- **Starts at $0** — <!-- br:models.free -->7<!-- /br:models.free --> open-weight models are free forever (incl. a 1M-context reasoner and two sub-second coders). Image turns route to paid models — see the free-tier note below.
 - **No accounts** — a wallet is generated locally, no signup
 - **No API keys** — your wallet signature IS authentication
 - **No model selection** — <!-- br:clawrouter.dimensions -->15<!-- /br:clawrouter.dimensions -->-dimension scoring + constraint-first ranking ([router-core](https://github.com/BlockRunAI/router-core)) picks the right model automatically
@@ -61,7 +61,7 @@ This is the stack that lets agents operate autonomously: **x402 + USDC + local r
 
 |                  | OpenRouter        | LiteLLM          | Martian           | Portkey           | **ClawRouter**                                                         |
 | ---------------- | ----------------- | ---------------- | ----------------- | ----------------- | ---------------------------------------------------------------------- |
-| **Models**       | 200+              | 100+             | Smart routing     | Gateway           | **<!-- br:models.chatVisible -->73<!-- /br:models.chatVisible -->**    |
+| **Models**       | 200+              | 100+             | Smart routing     | Gateway           | **<!-- br:models.chatVisible -->76<!-- /br:models.chatVisible -->**    |
 | **Free tier**    | Rate-limited      | BYO keys         | No                | No                | **<!-- br:models.free -->7<!-- /br:models.free --> models, no signup** |
 | **Routing**      | Manual selection  | Manual selection | Smart (closed)    | Observability     | **Smart (open source)**                                                |
 | **Auth**         | Account + API key | Your API keys    | Account + API key | Account + API key | **Wallet signature**                                                   |
@@ -107,7 +107,7 @@ openclaw gateway restart
 
 > **Using Claude Code?** Check out [BRCC](https://github.com/BlockRunAI/brcc) — it's purpose-built for Claude Code with the same smart routing and x402 payments.
 >
-> **Using NousResearch Hermes?** See [ClawRouter-Hermes](https://github.com/BlockRunAI/ClawRouter-Hermes) — a Python plugin that wires Hermes into the ClawRouter proxy. Same wallet, same <!-- br:models.chatVisible -->73<!-- /br:models.chatVisible --> models, same x402 USDC settlement on Base & Solana.
+> **Using NousResearch Hermes?** See [ClawRouter-Hermes](https://github.com/BlockRunAI/ClawRouter-Hermes) — a Python plugin that wires Hermes into the ClawRouter proxy. Same wallet, same <!-- br:models.chatVisible -->76<!-- /br:models.chatVisible --> models, same x402 USDC settlement on Base & Solana.
 
 No OpenClaw required. ClawRouter runs as a local proxy on port 8402.
 
@@ -191,7 +191,7 @@ Choose your routing strategy with `/model <profile>`:
 | Profile          | Strategy           | Savings                                                                            | Best For             |
 | ---------------- | ------------------ | ---------------------------------------------------------------------------------- | -------------------- |
 | `/model free`    | Free NVIDIA models | **100%**                                                                           | $0 balance, learning |
-| `/model auto`    | Balanced (default) | **<!-- br:savings.autoVsBaselinePct -->88<!-- /br:savings.autoVsBaselinePct -->%** | General use          |
+| `/model auto`    | Balanced (default) | **<!-- br:savings.autoVsBaselinePct -->84<!-- /br:savings.autoVsBaselinePct -->%** | General use          |
 | `/model eco`     | Cheapest possible  | **<!-- br:savings.ecoVsBaselinePct -->98<!-- /br:savings.ecoVsBaselinePct -->%**   | Maximum savings      |
 | `/model premium` | Best quality       | 0%                                                                                 | Mission-critical     |
 
@@ -229,7 +229,7 @@ Request → 15-dimension scorer → tier + task shape
 
 These are the curated primaries; every tier carries a benchmark-ordered fallback chain — full chains and per-profile ranking weights in [docs/routing-profiles.md](docs/routing-profiles.md). On router-core's frozen three-arm agent benchmark (τ-bench, BrowseComp, Terminal-Bench) the V3.4 policy (the last one benchmarked; V3.5 is a catalog refresh on the same scorer and weights) completed **57%** of tasks vs **49%** for the previous rules router, at **6.4%** lower cost per successful task — and spent **8.9%** of the tokens a pinned flagship would have.
 
-**<!-- br:savings.autoVsBaselinePct -->88<!-- /br:savings.autoVsBaselinePct -->% cheaper than pinning Claude Opus 5** for the same traffic, on `auto`; **<!-- br:savings.ecoVsBaselinePct -->98<!-- /br:savings.ecoVsBaselinePct -->%** on `eco`.
+**<!-- br:savings.autoVsBaselinePct -->84<!-- /br:savings.autoVsBaselinePct -->% cheaper than pinning Claude Opus 5** for the same traffic, on `auto`; **<!-- br:savings.ecoVsBaselinePct -->98<!-- /br:savings.ecoVsBaselinePct -->%** on `eco`.
 
 Not an "up to" figure. The baseline, the workload mix and the token ratio are
 published in [`savings-mix.json`](https://github.com/BlockRunAI/blockrun/blob/main/src/brand/savings-mix.json),
@@ -407,7 +407,7 @@ No Surf account, no API key — settles directly to Surf's Base treasury in USDC
 
 ## Models & Pricing
 
-<!-- br:models.chatVisible -->73<!-- /br:models.chatVisible --> models across 9 providers, one wallet. **<!-- br:models.free -->7<!-- /br:models.free --> models are $0 — paid models start at fractions of a cent.**
+<!-- br:models.chatVisible -->76<!-- /br:models.chatVisible --> models across 9 providers, one wallet. **<!-- br:models.free -->7<!-- /br:models.free --> models are $0 — paid models start at fractions of a cent.**
 
 > **💡 "Cost per request"** = estimated cost for a typical chat message (~500 input + 500 output tokens). Paid requests also carry a flat **$0.001/tx settlement fee** (covers on-chain gas; already included in the price the gateway quotes). Free models never pay it.
 
@@ -419,9 +419,9 @@ No Surf account, no API key — settles directly to Surf's Base treasury in USDC
 | free/nemotron-3-nano-30b                    |  **FREE** |   **FREE** |     **$0** | 131K    | fastest free model (~121 tok/s)              |
 | free/laguna-xs-2.1                          |  **FREE** |   **FREE** |     **$0** | 131K    | coding (Poolside, ~161 tok/s)                |
 | free/north-mini-code                        |  **FREE** |   **FREE** |     **$0** | 256K    | coding (Cohere, sub-second)                  |
-| free/nemotron-3-nano-omni-30b-a3b-reasoning |  **FREE** |   **FREE** |     **$0** | 256K    | reasoning, **vision** (text+img+video+audio) |
+| free/nemotron-3-nano-omni-30b-a3b-reasoning |  **FREE** |   **FREE** |     **$0** | 256K    | reasoning (text only — see note)             |
 | free/nemotron-3-ultra-550b                  |  **FREE** |   **FREE** |     **$0** | 1M      | largest free model (550B/55B MoE)            |
-| free/llama-3.2-11b-vision                   |  **FREE** |   **FREE** |     **$0** | 128K    | Meta Llama, image input                      |
+| free/llama-3.2-11b-vision                   |  **FREE** |   **FREE** |     **$0** | 128K    | Meta Llama (text only — see note)            |
 | qwen/qwen3.7-flash                          |     $0.03 |      $0.13 |    $0.0001 | 1M      | reasoning, tools (fastest Qwen tier)         |
 | openai/gpt-5-nano                           |     $0.05 |      $0.40 |    $0.0002 | 128K    | tools                                        |
 | openai/gpt-5.6-luna-pro                     |     $0.10 |      $0.60 |    $0.0004 | 1M      | reasoning, vision, agentic, tools            |
@@ -499,7 +499,9 @@ No Surf account, no API key — settles directly to Surf's Base treasury in USDC
 | openai/gpt-5.4-pro          |    $30.00 |    $180.00 |    $0.1050 | 400K    | reasoning, tools                  |
 | openai/gpt-5.5-pro          |    $30.00 |    $180.00 |    $0.1050 | 1M      | reasoning, vision, tools          |
 
-> **Free tier:** <!-- br:models.free -->7<!-- /br:models.free --> open-weight models cost nothing — `/model free` pins the free default (`nemotron-3.5-lightning`, with the other six as fallbacks), `/model eco` opens on it, or pick one directly (e.g., `/model nemotron-omni` for vision, `/model lightning` for 1M-context reasoning, `/model north-mini` or `/model laguna` for fast coding, `/model llama-vision` for Meta Llama).
+> **Free tier:** <!-- br:models.free -->7<!-- /br:models.free --> open-weight models cost nothing — `/model free` pins the free default (`nemotron-3.5-lightning`, with the other six as fallbacks), `/model eco` opens on it, or pick one directly (e.g., `/model lightning` for 1M-context reasoning, `/model north-mini` or `/model laguna` for fast coding, `/model nano-30b` for the fastest).
+>
+> **No free vision.** Two free models are catalogued as vision-capable and neither reliably is: on a 64×64 solid-colour probe `nemotron-3-nano-omni` was right 1 of 4 times on Base and answered "white" for red on Solana, and `llama-3.2-11b-vision` replied "I'm unable to see the image" 3 of 3 while answering text fine. Both return HTTP 200, so a wrong answer arrives with no error to branch on. ClawRouter therefore does not flag them for vision, and requests carrying an `image_url` route to a paid vision model instead.
 > **Best value:** `gpt-5-nano` and `gemini-2.5-flash-lite` deliver strong results at ~$0.0003/request.
 
 ---
@@ -669,7 +671,7 @@ npm test
 
 **The LLM router built for autonomous agents**
 
-You're here. <!-- br:models.chatVisible -->73<!-- /br:models.chatVisible --> models, local smart routing, x402 USDC payments — the only stack that lets agents operate independently.
+You're here. <!-- br:models.chatVisible -->76<!-- /br:models.chatVisible --> models, local smart routing, x402 USDC payments — the only stack that lets agents operate independently.
 
 `curl -fsSL https://blockrun.ai/ClawRouter-update | bash`
 
@@ -680,7 +682,7 @@ You're here. <!-- br:models.chatVisible -->73<!-- /br:models.chatVisible --> mod
 
 **BlockRun for Claude Code**
 
-Run Claude Code with <!-- br:models.chatVisible -->73<!-- /br:models.chatVisible --> models, no rate limits, no Anthropic account, no phone verification. Pay per request with USDC — your wallet is your identity.
+Run Claude Code with <!-- br:models.chatVisible -->76<!-- /br:models.chatVisible --> models, no rate limits, no Anthropic account, no phone verification. Pay per request with USDC — your wallet is your identity.
 
 `curl -fsSL https://blockrun.ai/brcc-install | bash`
 
@@ -693,7 +695,7 @@ Run Claude Code with <!-- br:models.chatVisible -->73<!-- /br:models.chatVisible
 
 **ClawRouter for NousResearch Hermes**
 
-Python plugin that wraps the ClawRouter proxy for `hermes-agent`. Same <!-- br:models.chatVisible -->73<!-- /br:models.chatVisible --> models, same x402 USDC payments on Base & Solana, native Hermes ergonomics.
+Python plugin that wraps the ClawRouter proxy for `hermes-agent`. Same <!-- br:models.chatVisible -->76<!-- /br:models.chatVisible --> models, same x402 USDC payments on Base & Solana, native Hermes ergonomics.
 
 `pip install hermes-plugin-clawrouter`
 
@@ -741,7 +743,7 @@ ClawRouter is an open-source (MIT licensed) smart LLM router built for autonomou
 
 ### How much can ClawRouter save on LLM costs?
 
-On the `auto` profile ClawRouter costs <!-- br:savings.autoVsBaselinePct -->88<!-- /br:savings.autoVsBaselinePct -->% less than pinning Claude Opus 5 for every request, and <!-- br:savings.ecoVsBaselinePct -->98<!-- /br:savings.ecoVsBaselinePct -->% less on `eco`. That is computed from a published workload mix rather than estimated — see [savings-mix.json](https://github.com/BlockRunAI/blockrun/blob/main/src/brand/savings-mix.json) for the baseline and assumptions. Actual savings depend on your workload — simple queries are routed to free models ($0/request), while complex tasks get premium models.
+On the `auto` profile ClawRouter costs <!-- br:savings.autoVsBaselinePct -->84<!-- /br:savings.autoVsBaselinePct -->% less than pinning Claude Opus 5 for every request, and <!-- br:savings.ecoVsBaselinePct -->98<!-- /br:savings.ecoVsBaselinePct -->% less on `eco`. That is computed from a published workload mix rather than estimated — see [savings-mix.json](https://github.com/BlockRunAI/blockrun/blob/main/src/brand/savings-mix.json) for the baseline and assumptions. Actual savings depend on your workload — simple queries are routed to free models ($0/request), while complex tasks get premium models.
 
 ### How does ClawRouter compare to OpenRouter?
 
@@ -757,7 +759,7 @@ ClawRouter works with any tool that makes OpenAI-compatible API calls — point
 
 ### Is ClawRouter free?
 
-ClawRouter itself is free and MIT licensed. You pay only for the LLM API calls routed through it — and several open-weight models (`nemotron-3.5-lightning`, `nemotron-3-nano-30b`, `laguna-xs-2.1`, `north-mini-code`, `nemotron-3-nano-omni-30b-a3b-reasoning`, `nemotron-3-ultra-550b`, `llama-3.2-11b-vision`) are completely free. Use `/model free` to smart-route across them, or pick any by name.
+ClawRouter itself is free and MIT licensed. You pay only for the LLM API calls routed through it — and several open-weight models (`nemotron-3.5-lightning`, `nemotron-3-nano-30b`, `laguna-xs-2.1`, `north-mini-code`, `nemotron-3-nano-omni-30b-a3b-reasoning`, `nemotron-3-ultra-550b`, `llama-3.2-11b-vision`) are completely free (text only). Use `/model free` to smart-route across them, or pick any by name.
 
 ---
 

package/dist/cli.js

@@ -37684,10 +37684,13 @@ var init_top_models = __esm({
       "zai/glm-5.1",
       "zai/glm-5-turbo",
       "zai/glm-5",
+      "xiaomi/mimo-v2.5",
       "minimax/minimax-m3",
       "minimax/minimax-m2.7",
       "moonshot/kimi-k3",
       "qwen/qwen3.7-max",
+      "qwen/qwen3.8-flash",
+      "deepseek/deepseek-v4-flash-vision-exp",
       "deepseek/deepseek-v4-pro",
       "deepseek/deepseek-chat",
       "deepseek/deepseek-reasoner",
@@ -37949,6 +37952,19 @@ var init_models = __esm({
       // `kimi` on K2.7. (`grok` WAS promoted to 4.5, but only after the cost
       // tradeoff was argued explicitly; there's no such case for qwen yet.)
       "qwen3.7-max": "qwen/qwen3.7-max",
+      // Qwen3.8 Flash — newer generation than the whole 3.7 line, and cheaper than
+      // the 3.7-plus tier it beats. Bare `qwen` stays UNBOUND (every other qwen*
+      // shorthand resolves to a free model, so binding the short name to a paid
+      // flagship would bill callers expecting free).
+      "qwen3.8-flash": "qwen/qwen3.8-flash",
+      "qwen3-8-flash": "qwen/qwen3.8-flash",
+      "qwen-vision": "qwen/qwen3.8-flash",
+      // DeepSeek's first image-capable SKU. Bare `deepseek` stays on deepseek-chat.
+      "deepseek-vision": "deepseek/deepseek-v4-flash-vision-exp",
+      "v4-flash-vision": "deepseek/deepseek-v4-flash-vision-exp",
+      // Xiaomi MiMo V2.5 — the natively multimodal SKU, distinct from mimo-v2.5-pro
+      // (text-only upstream). `mimo` stays on the Pro entry it has always named.
+      "mimo-vision": "xiaomi/mimo-v2.5",
       "qwen-3.7-max": "qwen/qwen3.7-max",
       "qwen3-7-max": "qwen/qwen3.7-max",
       // Plus/Flash tiers (2026-08-03) — explicit pins, same rule as Max.
@@ -37963,7 +37979,15 @@ var init_models = __esm({
       hunyuan: "tencent/hy3",
       mimo: "xiaomi/mimo-v2.5-pro",
       "mimo-v2.5-pro": "xiaomi/mimo-v2.5-pro",
-      "mimo-v2.5": "xiaomi/mimo-v2.5-pro",
+      // RETARGETED 2026-08-30, Pro -> the real thing. This key used to point at
+      // `xiaomi/mimo-v2.5-pro`, which was harmless while no model owned the name —
+      // but blockrun then listed an actual `xiaomi/mimo-v2.5`, a DIFFERENT and
+      // natively-multimodal SKU at $0.14/$0.28 against Pro's $0.435/$0.87. Leaving
+      // it would have billed 3x for the text-only model when the caller named the
+      // cheaper multimodal one. (The key itself is safe: it is not equal to the
+      // catalog id, so it shadows nothing — the rule that bans `opus-5`-style keys
+      // does not apply to a bare shorthand.)
+      "mimo-v2.5": "xiaomi/mimo-v2.5",
       xiaomi: "xiaomi/mimo-v2.5-pro",
       // Google
       // gemini-3-pro-preview delisted by Google 2026-06-06 — mirror the gateway
@@ -38138,6 +38162,10 @@ var init_models = __esm({
       // Vision-capable free models
       "nemotron-omni": "free/nemotron-3-nano-omni-30b-a3b-reasoning",
       "nano-omni": "free/nemotron-3-nano-omni-30b-a3b-reasoning",
+      // `vision-free` kept for backward compatibility ONLY — it still resolves to
+      // nano-omni, which is the strongest free model, but the free tier no longer
+      // claims working image input on either chain (see the catalog note). Do not
+      // advertise this alias as a way to get free vision.
       "vision-free": "free/nemotron-3-nano-omni-30b-a3b-reasoning",
       // Retired shorthand aliases redirect to live successors. The catch-all target
       // moved llama-4-maverick → gpt-oss-120b (2026-07-17) → step-3.7-flash
@@ -38403,8 +38431,8 @@ var init_models = __esm({
         id: "openai/gpt-5.6-terra-pro",
         name: "GPT-5.6 Terra Pro",
         version: "5.6",
-        inputPrice: 1,
-        outputPrice: 6,
+        inputPrice: 2,
+        outputPrice: 12,
         contextWindow: 105e4,
         maxOutput: 128e3,
         reasoning: true,
@@ -38416,8 +38444,8 @@ var init_models = __esm({
         id: "openai/gpt-5.6-luna-pro",
         name: "GPT-5.6 Luna Pro",
         version: "5.6",
-        inputPrice: 0.1,
-        outputPrice: 0.6,
+        inputPrice: 0.2,
+        outputPrice: 1.2,
         contextWindow: 105e4,
         maxOutput: 128e3,
         reasoning: true,
@@ -38950,8 +38978,8 @@ var init_models = __esm({
         id: "deepseek/deepseek-v4-pro",
         name: "DeepSeek V4 Pro",
         version: "4-pro",
-        inputPrice: 0.435,
-        outputPrice: 0.87,
+        inputPrice: 1.32,
+        outputPrice: 3.96,
         contextWindow: 1048576,
         maxOutput: 65536,
         reasoning: true,
@@ -39409,8 +39437,21 @@ var init_models = __esm({
         outputPrice: 0,
         contextWindow: 256e3,
         maxOutput: 16384,
-        reasoning: true,
-        vision: true
+        reasoning: true
+        // NO `vision: true` — 2026-08-31. Both catalogs advertise vision on this
+        // model and blockrun #448 cites an 8x8 PNG answering "Red", but a 64x64
+        // solid-red probe does not hold up: 1 of 4 correct on Base (the others
+        // "I'm not able to view the image" or leaked reasoning), and on sol it
+        // answered "white" twice, once with the response's own `model` field
+        // reading `nemotron-3-super-120b (fallback: ...nano-omni)` — the image is
+        // silently dropped and a text model answers.
+        //
+        // `vision: true` is what makes filterByVision() route real image turns
+        // here, so the flag does not merely describe the model, it aims traffic at
+        // it. HTTP 200 with a confident wrong colour is worse than no free vision:
+        // there is no error for a caller to branch on. Image turns go to paid
+        // vision models until a correctly-sized probe comes back right on both
+        // chains. Independently confirmed on Solana by the blockrun-sol owner.
       },
       // 2026-06-14: BlockRun re-featured these two as free flagships (catalog sweep).
       // Added to the auto-pick set behind gpt-oss to strengthen the mid/back of the
@@ -39532,8 +39573,11 @@ var init_models = __esm({
         outputPrice: 0,
         contextWindow: 131072,
         maxOutput: 16384,
-        reasoning: true,
-        vision: true
+        reasoning: true
+        // Vision flag dropped 2026-08-31 with the rest of the free tier: this id is
+        // 410 Gone and the gateway redirects it to nano-omni, whose image path does
+        // not work either (see above). A pinned caller sending an image would have
+        // been told, by the flag, that it would be seen.
       },
       // ── 2026-08-30 free-tier rebuild (blockrun #448) ─────────────────────────
       // NVIDIA retired FOUR of the five VISIBLE free models in one sweep. blockrun
@@ -39606,8 +39650,13 @@ var init_models = __esm({
         inputPrice: 0,
         outputPrice: 0,
         contextWindow: 128e3,
-        maxOutput: 16384,
-        vision: true
+        maxOutput: 16384
+        // NO `vision: true` despite the name and blockrun's `categories:
+        // ["chat","vision"]` — 2026-08-31, three consecutive 64x64 PNG probes came
+        // back "I'm unable to see the image" / "you haven't provided an image",
+        // while a plain-text control on the same id answered fine. The model is
+        // alive; the image path is not. Mirroring the catalog's claim here would
+        // have routed image turns to it. See the nano-omni note above.
       },
       {
         // Cohere North Mini Code, on OpenRouter's $0 pool — 607ms median, the
@@ -39637,6 +39686,57 @@ var init_models = __esm({
         contextWindow: 131072,
         maxOutput: 16384
       },
+      // ── 2026-08-30 paid catalog refresh (blockrun #449) ─────────────────────
+      // Three additions, each probe-verified upstream with a real completion AND a
+      // real image before listing. All three carry `vision: true` on that evidence
+      // — unlike the free tier, where the same claim did not survive a probe.
+      {
+        // Alibaba's 3.8 generation: 125B MoE, one tier above the whole 3.7 line and
+        // cheaper than the qwen3.7-plus ($0.32/$1.28) it outperforms.
+        // Vision took two probes to establish upstream: the first 400'd with
+        // `invalid_parameter_error` on an 8x8 PNG because the model requires >10px
+        // per side; 64x64 answered correctly. A single 400 is not a capability gap.
+        id: "qwen/qwen3.8-flash",
+        name: "Qwen3.8 Flash",
+        version: "3.8-flash",
+        inputPrice: 0.15,
+        outputPrice: 0.47,
+        contextWindow: 1e6,
+        maxOutput: 131072,
+        reasoning: true,
+        vision: true,
+        toolCalling: true
+      },
+      {
+        // The first DeepSeek SKU that takes images. Priced at DeepSeek's PEAK rate
+        // on purpose: they now split peak/off-peak and off-peak is half, so listing
+        // the lower number would sell under cost for seven hours every weekday.
+        id: "deepseek/deepseek-v4-flash-vision-exp",
+        name: "DeepSeek V4 Flash Vision",
+        version: "v4-flash-vision-exp",
+        inputPrice: 0.44,
+        outputPrice: 1.32,
+        contextWindow: 1048576,
+        maxOutput: 65536,
+        reasoning: true,
+        vision: true,
+        toolCalling: true
+      },
+      {
+        // NOT a cheaper mimo-v2.5-pro — a different, natively multimodal SKU. The
+        // Pro entry is text-only upstream while this one takes images, at a third
+        // of the price. Keep both.
+        id: "xiaomi/mimo-v2.5",
+        name: "Xiaomi MiMo V2.5",
+        version: "2.5",
+        inputPrice: 0.14,
+        outputPrice: 0.28,
+        contextWindow: 1048576,
+        maxOutput: 131072,
+        reasoning: true,
+        vision: true,
+        toolCalling: true
+      },
       // Z.AI GLM-5 Models
       {
         // Z.AI's flagship, live-probed by blockrun 2026-08-19 against api.z.ai
@@ -89923,7 +90023,36 @@ import { homedir as homedir7 } from "os";
 import { join as join10 } from "path";
 import { mkdir as mkdir4, writeFile as writeFile2, readFile as readFile2, stat as fsStat } from "fs/promises";
 import { readFileSync as readFileSync2, existsSync as existsSync2 } from "fs";
+async function loadGatewayCatalog(apiBase) {
+  try {
+    const controller = new AbortController();
+    const timer2 = setTimeout(() => controller.abort(), GATEWAY_CATALOG_TIMEOUT_MS);
+    const res = await fetch(`${apiBase}/v1/models`, { signal: controller.signal });
+    clearTimeout(timer2);
+    if (!res.ok) return;
+    const body = await res.json();
+    const ids = (body.data ?? []).map((m) => m.id).filter((id2) => !!id2);
+    if (ids.length === 0) return;
+    gatewayModelIds = new Set(ids);
+    const unserved = [...FREE_MODELS].filter((m) => !gatewayModelIds.has(toUpstreamModelId(m)));
+    if (unserved.length > 0) {
+      console.log(
+        `[ClawRouter] Free models not served by this gateway (skipped in the cascade): ${unserved.join(", ")}`
+      );
+    }
+  } catch {
+  }
+}
+function isServedByGateway(modelId) {
+  if (!gatewayModelIds) return true;
+  return gatewayModelIds.has(toUpstreamModelId(modelId));
+}
 function pickFreeModel(excludeList) {
+  for (const m of FREE_MODELS) {
+    if (excludeList?.has(m)) continue;
+    if (!isServedByGateway(m)) continue;
+    return m;
+  }
   for (const m of FREE_MODELS) {
     if (!excludeList?.has(m)) return m;
   }
@@ -90263,6 +90392,9 @@ function detectDegradedSuccessResponse(body) {
       return `degraded response: ${errorText.slice(0, 120)}`;
     }
     const choices = parsed.choices;
+    if (Array.isArray(choices) && choices.length === 0) {
+      return "degraded response: no choices returned";
+    }
     if (Array.isArray(choices) && choices.length > 0) {
       const choice = choices[0];
       const msg = choice.message ?? choice.delta;
@@ -90753,6 +90885,7 @@ async function startProxy(options) {
   } else if (paymentChain === "solana") {
     console.log(`[ClawRouter] Payment chain: Solana (${BLOCKRUN_SOLANA_API})`);
   }
+  void loadGatewayCatalog(apiBase);
   const listenPort = options.port ?? getProxyPort();
   const existingProxy = await checkExistingProxy(listenPort);
   if (existingProxy) {
@@ -93286,6 +93419,22 @@ data: [DONE]
     } else {
       modelsToTry = modelId ? [modelId] : [];
     }
+    if (modelsToTry.length > 0) {
+      const unservedFree = modelsToTry.filter((m) => FREE_MODELS.has(m) && !isServedByGateway(m));
+      if (unservedFree.length > 0) {
+        const servedFree = [...FREE_MODELS].filter(
+          (m) => isServedByGateway(m) && !excludeList?.has(m)
+        );
+        const replacement = servedFree[0];
+        if (replacement) {
+          modelsToTry = modelsToTry.map((m) => unservedFree.includes(m) ? replacement : m);
+          modelsToTry = modelsToTry.filter((m, i) => modelsToTry.indexOf(m) === i);
+          console.log(
+            `[ClawRouter] ${unservedFree.join(", ")} not served on this chain \u2014 using ${replacement}`
+          );
+        }
+      }
+    }
     if (!hasTools && routingDecision) {
       const freeFallback = pickFreeModel(excludeList);
       if (freeFallback && !modelsToTry.includes(freeFallback)) {
@@ -94166,7 +94315,7 @@ data: [DONE]
     });
   }
 }
-var paymentStore, BLOCKRUN_API, BLOCKRUN_SOLANA_API, IMAGE_DIR, AUDIO_DIR, VIDEO_DIR, AUTO_MODEL, ROUTING_PROFILES, FREE_MODELS, FREE_MODEL, FREE_UPSTREAM_OVERRIDES, MAX_MESSAGES, CONTEXT_LIMIT_KB, HEARTBEAT_INTERVAL_MS, BALANCE_CHECK_TIMEOUT_MS, DEFAULT_REQUEST_TIMEOUT_MS, PER_MODEL_TIMEOUT_MS, REASONING_MODEL_TIMEOUT_MS, REASONING_MODEL_IDS, MAX_FALLBACK_ATTEMPTS, HEALTH_CHECK_TIMEOUT_MS, RATE_LIMIT_COOLDOWN_MS, OVERLOAD_COOLDOWN_MS, PORT_RETRY_ATTEMPTS, PORT_RETRY_DELAY_MS, MODEL_BODY_READ_TIMEOUT_MS, ERROR_BODY_READ_TIMEOUT_MS, ClientDisconnectedError, rateLimitedModels, overloadedModels, perProviderErrors, BALANCE_CHECK_BUFFER, BALANCE_PREFLIGHT_OUTPUT_TOKEN_CAP, PROVIDER_ERROR_PATTERNS, DEGRADED_RESPONSE_PATTERNS, DEGRADED_LOOP_PATTERNS, VALID_ROLES, ROLE_MAPPINGS, VALID_TOOL_ID_PATTERN, KIMI_BLOCK_RE, KIMI_TOKEN_RE, THINKING_TAG_RE, THINKING_BLOCK_RE, BLOCKRUN_MODEL_BY_ID, IMAGE_PRICING, IMAGE_MODEL_IDS, IMAGE_MODEL_SIZES, IMAGE_MODEL_ALIASES, VIDEO_PRICING, PHONE_PRICING;
+var paymentStore, BLOCKRUN_API, BLOCKRUN_SOLANA_API, IMAGE_DIR, AUDIO_DIR, VIDEO_DIR, AUTO_MODEL, ROUTING_PROFILES, FREE_MODELS, gatewayModelIds, FREE_MODEL, FREE_UPSTREAM_OVERRIDES, MAX_MESSAGES, CONTEXT_LIMIT_KB, HEARTBEAT_INTERVAL_MS, BALANCE_CHECK_TIMEOUT_MS, GATEWAY_CATALOG_TIMEOUT_MS, DEFAULT_REQUEST_TIMEOUT_MS, PER_MODEL_TIMEOUT_MS, REASONING_MODEL_TIMEOUT_MS, REASONING_MODEL_IDS, MAX_FALLBACK_ATTEMPTS, HEALTH_CHECK_TIMEOUT_MS, RATE_LIMIT_COOLDOWN_MS, OVERLOAD_COOLDOWN_MS, PORT_RETRY_ATTEMPTS, PORT_RETRY_DELAY_MS, MODEL_BODY_READ_TIMEOUT_MS, ERROR_BODY_READ_TIMEOUT_MS, ClientDisconnectedError, rateLimitedModels, overloadedModels, perProviderErrors, BALANCE_CHECK_BUFFER, BALANCE_PREFLIGHT_OUTPUT_TOKEN_CAP, PROVIDER_ERROR_PATTERNS, DEGRADED_RESPONSE_PATTERNS, DEGRADED_LOOP_PATTERNS, VALID_ROLES, ROLE_MAPPINGS, VALID_TOOL_ID_PATTERN, KIMI_BLOCK_RE, KIMI_TOKEN_RE, THINKING_TAG_RE, THINKING_BLOCK_RE, BLOCKRUN_MODEL_BY_ID, IMAGE_PRICING, IMAGE_MODEL_IDS, IMAGE_MODEL_SIZES, IMAGE_MODEL_ALIASES, VIDEO_PRICING, PHONE_PRICING;
 var init_proxy = __esm({
   "src/proxy.ts"() {
     "use strict";
@@ -94241,6 +94390,7 @@ var init_proxy = __esm({
     CONTEXT_LIMIT_KB = 5120;
     HEARTBEAT_INTERVAL_MS = 2e3;
     BALANCE_CHECK_TIMEOUT_MS = 2500;
+    GATEWAY_CATALOG_TIMEOUT_MS = 5e3;
     DEFAULT_REQUEST_TIMEOUT_MS = 3e5;
     PER_MODEL_TIMEOUT_MS = 6e4;
     REASONING_MODEL_TIMEOUT_MS = 18e4;
@@ -226642,7 +226792,8 @@ import {
   readdirSync,
   mkdirSync as mkdirSync3,
   copyFileSync,
-  renameSync as renameSync2
+  renameSync as renameSync2,
+  unlinkSync
 } from "fs";
 import { readFile as readFileAsync } from "fs/promises";
 import { homedir as homedir13 } from "os";
@@ -226987,6 +227138,15 @@ function injectAuthProfile(logger48) {
     for (const agentId of agents) {
       const authDir = join16(agentsDir, agentId, "agent");
       const authPath = join16(authDir, "auth-profiles.json");
+      const sqlitePath = join16(authDir, "openclaw-agent.sqlite");
+      if (existsSync4(sqlitePath)) {
+        removeInjectedAuthPlaceholder(authPath, logger48, agentId);
+        continue;
+      }
+      if (agentId === "main") {
+        removeInjectedAuthPlaceholder(authPath, logger48, agentId);
+        continue;
+      }
       if (!existsSync4(authDir)) {
         try {
           mkdirSync3(authDir, { recursive: true });
@@ -227029,6 +227189,25 @@ function injectAuthProfile(logger48) {
     logger48.info(`Auth injection failed: ${err instanceof Error ? err.message : String(err)}`);
   }
 }
+function removeInjectedAuthPlaceholder(authPath, logger48, agentId) {
+  try {
+    if (!existsSync4(authPath)) return;
+    const parsed = JSON.parse(readTextFileSync(authPath));
+    const profiles = parsed?.profiles;
+    if (!profiles || typeof profiles !== "object" || Array.isArray(profiles)) return;
+    const keys = Object.keys(profiles);
+    if (keys.length !== 1 || keys[0] !== "blockrun:default") return;
+    const entry = profiles["blockrun:default"];
+    if (!entry || typeof entry !== "object") return;
+    const profile = entry;
+    if (profile.type !== "api_key" || profile.provider !== "blockrun" || profile.key !== "x402-proxy-handles-auth") {
+      return;
+    }
+    unlinkSync(authPath);
+    logger48.info(`Removed legacy BlockRun auth placeholder for agent: ${agentId}`);
+  } catch {
+  }
+}
 function clearDeferredProxyStartTimer(proc = process) {
   if (!proc.__clawrouterDeferredStartTimer) return false;
   clearTimeout(proc.__clawrouterDeferredStartTimer);

package/dist/index.d.ts

@@ -1692,8 +1692,21 @@ declare function syncAgentModelCache(logger: {
 }): void;
 /**
  * Inject dummy auth profile for BlockRun into agent auth stores.
- * OpenClaw's agent system looks for auth credentials even if provider has auth: [].
- * We inject a placeholder so the lookup succeeds (proxy handles real auth internally).
+ *
+ * The legacy ``auth-profiles.json`` write is now deliberately narrow:
+ *
+ * - Wherever ``openclaw-agent.sqlite`` exists, the SQLite auth store is
+ *   authoritative. Writing the legacy JSON beside it is at best ignored, at
+ *   worst a failed-closed migration trigger: since OpenClaw 2026.8.1 a
+ *   leftover legacy file beside a store that holds no profiles fails auth
+ *   migration for the whole agent fleet. So we never write there, and we
+ *   clean up the placeholder we previously injected.
+ * - The shared auth-owner directory (``main``) is managed by OpenClaw
+ *   itself; a placeholder written there can shadow that state. The
+ *   provider's real auth comes from the x402 proxy (and the apiKey
+ *   injectModelsConfig writes into openclaw.json), so nothing is lost.
+ * - Only on very old installs with no SQLite store at all do we keep the
+ *   original JSON bootstrap, which those releases import.
  */
 declare function injectAuthProfile(logger: {
     info: (msg: string) => void;

package/dist/index.js

@@ -87,10 +87,13 @@ var init_top_models = __esm({
       "zai/glm-5.1",
       "zai/glm-5-turbo",
       "zai/glm-5",
+      "xiaomi/mimo-v2.5",
       "minimax/minimax-m3",
       "minimax/minimax-m2.7",
       "moonshot/kimi-k3",
       "qwen/qwen3.7-max",
+      "qwen/qwen3.8-flash",
+      "deepseek/deepseek-v4-flash-vision-exp",
       "deepseek/deepseek-v4-pro",
       "deepseek/deepseek-chat",
       "deepseek/deepseek-reasoner",
@@ -352,6 +355,19 @@ var init_models = __esm({
       // `kimi` on K2.7. (`grok` WAS promoted to 4.5, but only after the cost
       // tradeoff was argued explicitly; there's no such case for qwen yet.)
       "qwen3.7-max": "qwen/qwen3.7-max",
+      // Qwen3.8 Flash — newer generation than the whole 3.7 line, and cheaper than
+      // the 3.7-plus tier it beats. Bare `qwen` stays UNBOUND (every other qwen*
+      // shorthand resolves to a free model, so binding the short name to a paid
+      // flagship would bill callers expecting free).
+      "qwen3.8-flash": "qwen/qwen3.8-flash",
+      "qwen3-8-flash": "qwen/qwen3.8-flash",
+      "qwen-vision": "qwen/qwen3.8-flash",
+      // DeepSeek's first image-capable SKU. Bare `deepseek` stays on deepseek-chat.
+      "deepseek-vision": "deepseek/deepseek-v4-flash-vision-exp",
+      "v4-flash-vision": "deepseek/deepseek-v4-flash-vision-exp",
+      // Xiaomi MiMo V2.5 — the natively multimodal SKU, distinct from mimo-v2.5-pro
+      // (text-only upstream). `mimo` stays on the Pro entry it has always named.
+      "mimo-vision": "xiaomi/mimo-v2.5",
       "qwen-3.7-max": "qwen/qwen3.7-max",
       "qwen3-7-max": "qwen/qwen3.7-max",
       // Plus/Flash tiers (2026-08-03) — explicit pins, same rule as Max.
@@ -366,7 +382,15 @@ var init_models = __esm({
       hunyuan: "tencent/hy3",
       mimo: "xiaomi/mimo-v2.5-pro",
       "mimo-v2.5-pro": "xiaomi/mimo-v2.5-pro",
-      "mimo-v2.5": "xiaomi/mimo-v2.5-pro",
+      // RETARGETED 2026-08-30, Pro -> the real thing. This key used to point at
+      // `xiaomi/mimo-v2.5-pro`, which was harmless while no model owned the name —
+      // but blockrun then listed an actual `xiaomi/mimo-v2.5`, a DIFFERENT and
+      // natively-multimodal SKU at $0.14/$0.28 against Pro's $0.435/$0.87. Leaving
+      // it would have billed 3x for the text-only model when the caller named the
+      // cheaper multimodal one. (The key itself is safe: it is not equal to the
+      // catalog id, so it shadows nothing — the rule that bans `opus-5`-style keys
+      // does not apply to a bare shorthand.)
+      "mimo-v2.5": "xiaomi/mimo-v2.5",
       xiaomi: "xiaomi/mimo-v2.5-pro",
       // Google
       // gemini-3-pro-preview delisted by Google 2026-06-06 — mirror the gateway
@@ -541,6 +565,10 @@ var init_models = __esm({
       // Vision-capable free models
       "nemotron-omni": "free/nemotron-3-nano-omni-30b-a3b-reasoning",
       "nano-omni": "free/nemotron-3-nano-omni-30b-a3b-reasoning",
+      // `vision-free` kept for backward compatibility ONLY — it still resolves to
+      // nano-omni, which is the strongest free model, but the free tier no longer
+      // claims working image input on either chain (see the catalog note). Do not
+      // advertise this alias as a way to get free vision.
       "vision-free": "free/nemotron-3-nano-omni-30b-a3b-reasoning",
       // Retired shorthand aliases redirect to live successors. The catch-all target
       // moved llama-4-maverick → gpt-oss-120b (2026-07-17) → step-3.7-flash
@@ -806,8 +834,8 @@ var init_models = __esm({
         id: "openai/gpt-5.6-terra-pro",
         name: "GPT-5.6 Terra Pro",
         version: "5.6",
-        inputPrice: 1,
-        outputPrice: 6,
+        inputPrice: 2,
+        outputPrice: 12,
         contextWindow: 105e4,
         maxOutput: 128e3,
         reasoning: true,
@@ -819,8 +847,8 @@ var init_models = __esm({
         id: "openai/gpt-5.6-luna-pro",
         name: "GPT-5.6 Luna Pro",
         version: "5.6",
-        inputPrice: 0.1,
-        outputPrice: 0.6,
+        inputPrice: 0.2,
+        outputPrice: 1.2,
         contextWindow: 105e4,
         maxOutput: 128e3,
         reasoning: true,
@@ -1353,8 +1381,8 @@ var init_models = __esm({
         id: "deepseek/deepseek-v4-pro",
         name: "DeepSeek V4 Pro",
         version: "4-pro",
-        inputPrice: 0.435,
-        outputPrice: 0.87,
+        inputPrice: 1.32,
+        outputPrice: 3.96,
         contextWindow: 1048576,
         maxOutput: 65536,
         reasoning: true,
@@ -1812,8 +1840,21 @@ var init_models = __esm({
         outputPrice: 0,
         contextWindow: 256e3,
         maxOutput: 16384,
-        reasoning: true,
-        vision: true
+        reasoning: true
+        // NO `vision: true` — 2026-08-31. Both catalogs advertise vision on this
+        // model and blockrun #448 cites an 8x8 PNG answering "Red", but a 64x64
+        // solid-red probe does not hold up: 1 of 4 correct on Base (the others
+        // "I'm not able to view the image" or leaked reasoning), and on sol it
+        // answered "white" twice, once with the response's own `model` field
+        // reading `nemotron-3-super-120b (fallback: ...nano-omni)` — the image is
+        // silently dropped and a text model answers.
+        //
+        // `vision: true` is what makes filterByVision() route real image turns
+        // here, so the flag does not merely describe the model, it aims traffic at
+        // it. HTTP 200 with a confident wrong colour is worse than no free vision:
+        // there is no error for a caller to branch on. Image turns go to paid
+        // vision models until a correctly-sized probe comes back right on both
+        // chains. Independently confirmed on Solana by the blockrun-sol owner.
       },
       // 2026-06-14: BlockRun re-featured these two as free flagships (catalog sweep).
       // Added to the auto-pick set behind gpt-oss to strengthen the mid/back of the
@@ -1935,8 +1976,11 @@ var init_models = __esm({
         outputPrice: 0,
         contextWindow: 131072,
         maxOutput: 16384,
-        reasoning: true,
-        vision: true
+        reasoning: true
+        // Vision flag dropped 2026-08-31 with the rest of the free tier: this id is
+        // 410 Gone and the gateway redirects it to nano-omni, whose image path does
+        // not work either (see above). A pinned caller sending an image would have
+        // been told, by the flag, that it would be seen.
       },
       // ── 2026-08-30 free-tier rebuild (blockrun #448) ─────────────────────────
       // NVIDIA retired FOUR of the five VISIBLE free models in one sweep. blockrun
@@ -2009,8 +2053,13 @@ var init_models = __esm({
         inputPrice: 0,
         outputPrice: 0,
         contextWindow: 128e3,
-        maxOutput: 16384,
-        vision: true
+        maxOutput: 16384
+        // NO `vision: true` despite the name and blockrun's `categories:
+        // ["chat","vision"]` — 2026-08-31, three consecutive 64x64 PNG probes came
+        // back "I'm unable to see the image" / "you haven't provided an image",
+        // while a plain-text control on the same id answered fine. The model is
+        // alive; the image path is not. Mirroring the catalog's claim here would
+        // have routed image turns to it. See the nano-omni note above.
       },
       {
         // Cohere North Mini Code, on OpenRouter's $0 pool — 607ms median, the
@@ -2040,6 +2089,57 @@ var init_models = __esm({
         contextWindow: 131072,
         maxOutput: 16384
       },
+      // ── 2026-08-30 paid catalog refresh (blockrun #449) ─────────────────────
+      // Three additions, each probe-verified upstream with a real completion AND a
+      // real image before listing. All three carry `vision: true` on that evidence
+      // — unlike the free tier, where the same claim did not survive a probe.
+      {
+        // Alibaba's 3.8 generation: 125B MoE, one tier above the whole 3.7 line and
+        // cheaper than the qwen3.7-plus ($0.32/$1.28) it outperforms.
+        // Vision took two probes to establish upstream: the first 400'd with
+        // `invalid_parameter_error` on an 8x8 PNG because the model requires >10px
+        // per side; 64x64 answered correctly. A single 400 is not a capability gap.
+        id: "qwen/qwen3.8-flash",
+        name: "Qwen3.8 Flash",
+        version: "3.8-flash",
+        inputPrice: 0.15,
+        outputPrice: 0.47,
+        contextWindow: 1e6,
+        maxOutput: 131072,
+        reasoning: true,
+        vision: true,
+        toolCalling: true
+      },
+      {
+        // The first DeepSeek SKU that takes images. Priced at DeepSeek's PEAK rate
+        // on purpose: they now split peak/off-peak and off-peak is half, so listing
+        // the lower number would sell under cost for seven hours every weekday.
+        id: "deepseek/deepseek-v4-flash-vision-exp",
+        name: "DeepSeek V4 Flash Vision",
+        version: "v4-flash-vision-exp",
+        inputPrice: 0.44,
+        outputPrice: 1.32,
+        contextWindow: 1048576,
+        maxOutput: 65536,
+        reasoning: true,
+        vision: true,
+        toolCalling: true
+      },
+      {
+        // NOT a cheaper mimo-v2.5-pro — a different, natively multimodal SKU. The
+        // Pro entry is text-only upstream while this one takes images, at a third
+        // of the price. Keep both.
+        id: "xiaomi/mimo-v2.5",
+        name: "Xiaomi MiMo V2.5",
+        version: "2.5",
+        inputPrice: 0.14,
+        outputPrice: 0.28,
+        contextWindow: 1048576,
+        maxOutput: 131072,
+        reasoning: true,
+        vision: true,
+        toolCalling: true
+      },
       // Z.AI GLM-5 Models
       {
         // Z.AI's flagship, live-probed by blockrun 2026-08-19 against api.z.ai
@@ -89743,7 +89843,36 @@ import { homedir as homedir7 } from "os";
 import { join as join10 } from "path";
 import { mkdir as mkdir4, writeFile as writeFile2, readFile as readFile2, stat as fsStat } from "fs/promises";
 import { readFileSync as readFileSync2, existsSync as existsSync2 } from "fs";
+async function loadGatewayCatalog(apiBase) {
+  try {
+    const controller = new AbortController();
+    const timer2 = setTimeout(() => controller.abort(), GATEWAY_CATALOG_TIMEOUT_MS);
+    const res = await fetch(`${apiBase}/v1/models`, { signal: controller.signal });
+    clearTimeout(timer2);
+    if (!res.ok) return;
+    const body = await res.json();
+    const ids = (body.data ?? []).map((m) => m.id).filter((id2) => !!id2);
+    if (ids.length === 0) return;
+    gatewayModelIds = new Set(ids);
+    const unserved = [...FREE_MODELS].filter((m) => !gatewayModelIds.has(toUpstreamModelId(m)));
+    if (unserved.length > 0) {
+      console.log(
+        `[ClawRouter] Free models not served by this gateway (skipped in the cascade): ${unserved.join(", ")}`
+      );
+    }
+  } catch {
+  }
+}
+function isServedByGateway(modelId) {
+  if (!gatewayModelIds) return true;
+  return gatewayModelIds.has(toUpstreamModelId(modelId));
+}
 function pickFreeModel(excludeList) {
+  for (const m of FREE_MODELS) {
+    if (excludeList?.has(m)) continue;
+    if (!isServedByGateway(m)) continue;
+    return m;
+  }
   for (const m of FREE_MODELS) {
     if (!excludeList?.has(m)) return m;
   }
@@ -90083,6 +90212,9 @@ function detectDegradedSuccessResponse(body) {
       return `degraded response: ${errorText.slice(0, 120)}`;
     }
     const choices = parsed.choices;
+    if (Array.isArray(choices) && choices.length === 0) {
+      return "degraded response: no choices returned";
+    }
     if (Array.isArray(choices) && choices.length > 0) {
       const choice = choices[0];
       const msg = choice.message ?? choice.delta;
@@ -90573,6 +90705,7 @@ async function startProxy(options) {
   } else if (paymentChain === "solana") {
     console.log(`[ClawRouter] Payment chain: Solana (${BLOCKRUN_SOLANA_API})`);
   }
+  void loadGatewayCatalog(apiBase);
   const listenPort = options.port ?? getProxyPort();
   const existingProxy = await checkExistingProxy(listenPort);
   if (existingProxy) {
@@ -93106,6 +93239,22 @@ data: [DONE]
     } else {
       modelsToTry = modelId ? [modelId] : [];
     }
+    if (modelsToTry.length > 0) {
+      const unservedFree = modelsToTry.filter((m) => FREE_MODELS.has(m) && !isServedByGateway(m));
+      if (unservedFree.length > 0) {
+        const servedFree = [...FREE_MODELS].filter(
+          (m) => isServedByGateway(m) && !excludeList?.has(m)
+        );
+        const replacement = servedFree[0];
+        if (replacement) {
+          modelsToTry = modelsToTry.map((m) => unservedFree.includes(m) ? replacement : m);
+          modelsToTry = modelsToTry.filter((m, i) => modelsToTry.indexOf(m) === i);
+          console.log(
+            `[ClawRouter] ${unservedFree.join(", ")} not served on this chain \u2014 using ${replacement}`
+          );
+        }
+      }
+    }
     if (!hasTools && routingDecision) {
       const freeFallback = pickFreeModel(excludeList);
       if (freeFallback && !modelsToTry.includes(freeFallback)) {
@@ -93986,7 +94135,7 @@ data: [DONE]
     });
   }
 }
-var paymentStore, BLOCKRUN_API, BLOCKRUN_SOLANA_API, IMAGE_DIR, AUDIO_DIR, VIDEO_DIR, AUTO_MODEL, ROUTING_PROFILES, FREE_MODELS, FREE_MODEL, FREE_UPSTREAM_OVERRIDES, MAX_MESSAGES, CONTEXT_LIMIT_KB, HEARTBEAT_INTERVAL_MS, BALANCE_CHECK_TIMEOUT_MS, DEFAULT_REQUEST_TIMEOUT_MS, PER_MODEL_TIMEOUT_MS, REASONING_MODEL_TIMEOUT_MS, REASONING_MODEL_IDS, MAX_FALLBACK_ATTEMPTS, HEALTH_CHECK_TIMEOUT_MS, RATE_LIMIT_COOLDOWN_MS, OVERLOAD_COOLDOWN_MS, PORT_RETRY_ATTEMPTS, PORT_RETRY_DELAY_MS, MODEL_BODY_READ_TIMEOUT_MS, ERROR_BODY_READ_TIMEOUT_MS, ClientDisconnectedError, rateLimitedModels, overloadedModels, perProviderErrors, BALANCE_CHECK_BUFFER, BALANCE_PREFLIGHT_OUTPUT_TOKEN_CAP, PROVIDER_ERROR_PATTERNS, DEGRADED_RESPONSE_PATTERNS, DEGRADED_LOOP_PATTERNS, VALID_ROLES, ROLE_MAPPINGS, VALID_TOOL_ID_PATTERN, KIMI_BLOCK_RE, KIMI_TOKEN_RE, THINKING_TAG_RE, THINKING_BLOCK_RE, BLOCKRUN_MODEL_BY_ID, IMAGE_PRICING, IMAGE_MODEL_IDS, IMAGE_MODEL_SIZES, IMAGE_MODEL_ALIASES, VIDEO_PRICING, PHONE_PRICING;
+var paymentStore, BLOCKRUN_API, BLOCKRUN_SOLANA_API, IMAGE_DIR, AUDIO_DIR, VIDEO_DIR, AUTO_MODEL, ROUTING_PROFILES, FREE_MODELS, gatewayModelIds, FREE_MODEL, FREE_UPSTREAM_OVERRIDES, MAX_MESSAGES, CONTEXT_LIMIT_KB, HEARTBEAT_INTERVAL_MS, BALANCE_CHECK_TIMEOUT_MS, GATEWAY_CATALOG_TIMEOUT_MS, DEFAULT_REQUEST_TIMEOUT_MS, PER_MODEL_TIMEOUT_MS, REASONING_MODEL_TIMEOUT_MS, REASONING_MODEL_IDS, MAX_FALLBACK_ATTEMPTS, HEALTH_CHECK_TIMEOUT_MS, RATE_LIMIT_COOLDOWN_MS, OVERLOAD_COOLDOWN_MS, PORT_RETRY_ATTEMPTS, PORT_RETRY_DELAY_MS, MODEL_BODY_READ_TIMEOUT_MS, ERROR_BODY_READ_TIMEOUT_MS, ClientDisconnectedError, rateLimitedModels, overloadedModels, perProviderErrors, BALANCE_CHECK_BUFFER, BALANCE_PREFLIGHT_OUTPUT_TOKEN_CAP, PROVIDER_ERROR_PATTERNS, DEGRADED_RESPONSE_PATTERNS, DEGRADED_LOOP_PATTERNS, VALID_ROLES, ROLE_MAPPINGS, VALID_TOOL_ID_PATTERN, KIMI_BLOCK_RE, KIMI_TOKEN_RE, THINKING_TAG_RE, THINKING_BLOCK_RE, BLOCKRUN_MODEL_BY_ID, IMAGE_PRICING, IMAGE_MODEL_IDS, IMAGE_MODEL_SIZES, IMAGE_MODEL_ALIASES, VIDEO_PRICING, PHONE_PRICING;
 var init_proxy = __esm({
   "src/proxy.ts"() {
     "use strict";
@@ -94061,6 +94210,7 @@ var init_proxy = __esm({
     CONTEXT_LIMIT_KB = 5120;
     HEARTBEAT_INTERVAL_MS = 2e3;
     BALANCE_CHECK_TIMEOUT_MS = 2500;
+    GATEWAY_CATALOG_TIMEOUT_MS = 5e3;
     DEFAULT_REQUEST_TIMEOUT_MS = 3e5;
     PER_MODEL_TIMEOUT_MS = 6e4;
     REASONING_MODEL_TIMEOUT_MS = 18e4;
@@ -226389,7 +226539,8 @@ import {
   readdirSync,
   mkdirSync as mkdirSync3,
   copyFileSync,
-  renameSync as renameSync2
+  renameSync as renameSync2,
+  unlinkSync
 } from "fs";
 import { readFile as readFileAsync } from "fs/promises";
 import { homedir as homedir13 } from "os";
@@ -226734,6 +226885,15 @@ function injectAuthProfile(logger48) {
     for (const agentId of agents) {
       const authDir = join16(agentsDir, agentId, "agent");
       const authPath = join16(authDir, "auth-profiles.json");
+      const sqlitePath = join16(authDir, "openclaw-agent.sqlite");
+      if (existsSync3(sqlitePath)) {
+        removeInjectedAuthPlaceholder(authPath, logger48, agentId);
+        continue;
+      }
+      if (agentId === "main") {
+        removeInjectedAuthPlaceholder(authPath, logger48, agentId);
+        continue;
+      }
       if (!existsSync3(authDir)) {
         try {
           mkdirSync3(authDir, { recursive: true });
@@ -226776,6 +226936,25 @@ function injectAuthProfile(logger48) {
     logger48.info(`Auth injection failed: ${err instanceof Error ? err.message : String(err)}`);
   }
 }
+function removeInjectedAuthPlaceholder(authPath, logger48, agentId) {
+  try {
+    if (!existsSync3(authPath)) return;
+    const parsed = JSON.parse(readTextFileSync(authPath));
+    const profiles = parsed?.profiles;
+    if (!profiles || typeof profiles !== "object" || Array.isArray(profiles)) return;
+    const keys = Object.keys(profiles);
+    if (keys.length !== 1 || keys[0] !== "blockrun:default") return;
+    const entry = profiles["blockrun:default"];
+    if (!entry || typeof entry !== "object") return;
+    const profile = entry;
+    if (profile.type !== "api_key" || profile.provider !== "blockrun" || profile.key !== "x402-proxy-handles-auth") {
+      return;
+    }
+    unlinkSync(authPath);
+    logger48.info(`Removed legacy BlockRun auth placeholder for agent: ${agentId}`);
+  } catch {
+  }
+}
 function clearDeferredProxyStartTimer(proc = process) {
   if (!proc.__clawrouterDeferredStartTimer) return false;
   clearTimeout(proc.__clawrouterDeferredStartTimer);

package/docs/9-free-ai-models-zero-cost-blockrun.md

@@ -32,8 +32,8 @@ Through [ClawRouter](https://github.com/BlockRunAI/ClawRouter) — BlockRun's lo
 | **Nemotron 3.5 Lightning**   | 1M      | ✅        | The free default — thinking-mode reasoning |
 | **Nemotron 3 Nano 30B**      | 131K    | ✅        | Fastest free model (~121 tok/s)            |
 | **Nemotron 3 Ultra 550B**    | 1M      | ✅        | Largest free model — 550B / 55B active MoE |
-| **Nemotron 3 Nano Omni 30B** | 256K    | ✅        | Vision — text + image + video + audio      |
-| **Llama 3.2 11B Vision**     | 128K    | —         | Meta Llama with image input                |
+| **Nemotron 3 Nano Omni 30B** | 256K    | ✅        | Strong generalist (text only in practice)  |
+| **Llama 3.2 11B Vision**     | 128K    | —         | Meta Llama (text only in practice)         |
 | **Cohere North Mini Code**   | 256K    | ✅        | Coding, sub-second responses               |
 | **Poolside Laguna XS 2.1**   | 131K    | —         | Coding, ~161 tok/s                         |
 
@@ -111,17 +111,17 @@ The **fastest free model in the catalog** — around 121 tokens/second on realis
 
 **Best for:** Complex analysis and deep reasoning where you would otherwise pay.
 
-### Nemotron 3 Nano Omni 30B — The Only Free Vision Model
+### Nemotron 3 Nano Omni 30B — Strong, and Text-Only in Practice
 
-31B / 3.2B active MoE, 256K context, and the **only vision-capable model in the free tier**. It accepts images, video up to two minutes, and audio up to an hour. ChartQA 90.3, DocVQA 95.6, MMMU 70.8.
+31B / 3.2B active MoE, 256K context. Both gateways catalogue it as vision-capable and its benchmark card is real (ChartQA 90.3, DocVQA 95.6, MMMU 70.8) — but **the image path does not work through either gateway**, so ClawRouter treats it as text-only. See "There Is No Working Free Vision" below.
 
-**Best for:** Screenshots, charts, documents, and any multimodal input at zero cost.
+**Best for:** General reasoning at 256K context, for free.
 
 ### Llama 3.2 11B Vision — The Free Llama
 
-Meta's Llama 3.2 11B with image input and a 128K window. It is older than everything else here, and that is the point: a 12-model sweep of what NVIDIA still serves free found this was the only Llama that actually finishes a real completion.
+Meta's Llama 3.2 11B with a 128K window. It is older than everything else here, and that is the point: a 12-model sweep of what NVIDIA still serves free found this was the only Llama that actually finishes a real completion. The name promises image input; the gateway does not deliver it (see below).
 
-**Best for:** Anyone who needs a Llama specifically, plus lightweight image understanding.
+**Best for:** Anyone who needs a Llama specifically.
 
 ### Cohere North Mini Code — The Fast Coder
 
@@ -178,8 +178,8 @@ If you're using Claude Code, one command switches you to any free model:
 /model free              → Nemotron 3.5 Lightning (1M ctx, the default)
 /model nano-30b          → Nemotron 3 Nano 30B (fastest)
 /model ultra-550b        → Nemotron 3 Ultra 550B (largest)
-/model vision-free       → Nemotron 3 Nano Omni (images, video, audio)
-/model llama-vision      → Llama 3.2 11B Vision
+/model vision-free       → Nemotron 3 Nano Omni (text only, see below)
+/model llama-vision      → Llama 3.2 11B Vision (text only, see below)
 /model north-mini        → Cohere North Mini Code
 /model laguna            → Poolside Laguna XS 2.1
 ```
@@ -200,7 +200,13 @@ The gateway will return structured tool calls from these models, but ClawRouter
 
 Five of the seven are reasoning-capable and handle most tasks well. On the hardest problems — competition-level math, formal proofs, deep multi-step planning — they don't match Claude Opus 5 or Sonnet 5. That's why ClawRouter's REASONING tier doesn't use free models.
 
-### 3. Free Hosting Is Volatile
+### 3. There Is No Working Free Vision
+
+Two of the seven are catalogued as vision-capable, and neither survives a real probe. On a 64×64 solid-red PNG, `nemotron-3-nano-omni` answered correctly 1 time in 4 on Base and returned "white" on Solana — where the response's own `model` field revealed a silent fallback to a text model. `llama-3.2-11b-vision` replied "I'm unable to see the image" on 3 of 3 attempts while answering plain text fine. Every one of those failures is an **HTTP 200**: the image is dropped and a confident wrong answer comes back with no error to branch on.
+
+ClawRouter therefore ships no `vision` flag on any free model, so requests carrying an image route to a paid vision model instead. The general lesson is worth more than the specific finding: **a catalog's capability list is a claim, not a measurement** — and one passing sample is not a measurement either. The first probe here returned the right colour; it took four to see that was luck.
+
+### 4. Free Hosting Is Volatile
 
 This is the real limitation, and it is worth more than the other two. On 2026-08-30 NVIDIA retired four of the five visible free models in one sweep, and none of the nine models this article originally listed is still in the tier. Free capacity comes from whatever a provider is willing to give away this quarter. Pin a free model if you like — but build so that losing it costs you a config line, not a rewrite. That is the entire argument for putting a router in front of them.
 
@@ -217,7 +223,7 @@ Quick chat, formatting    → Nemotron 3 Nano 30B (fastest)
 Code generation           → North Mini Code or Laguna XS 2.1
 Reasoning required        → Nemotron 3.5 Lightning (1M ctx)
 Hardest free reasoning    → Nemotron 3 Ultra 550B
-Images, video, audio      → Nemotron 3 Nano Omni
+Long-context reasoning    → Nemotron 3.5 Lightning (1M)
 A Llama specifically      → Llama 3.2 11B Vision
 ```
 

package/docs/anthropic-cost-savings.md

@@ -50,7 +50,7 @@ This is where you're paying for real value:
 
 ## The Solution: ClawRouter
 
-[ClawRouter](https://github.com/BlockRunAI/ClawRouter) is an open-source local proxy that sits between your app and <!-- br:models.chatVisible -->73<!-- /br:models.chatVisible --> AI models. It saves you money in three ways: **smart routing**, **token optimization**, and **response caching**.
+[ClawRouter](https://github.com/BlockRunAI/ClawRouter) is an open-source local proxy that sits between your app and <!-- br:models.chatVisible -->76<!-- /br:models.chatVisible --> AI models. It saves you money in three ways: **smart routing**, **token optimization**, and **response caching**.
 
 ```
 ┌─────────────┐     ┌──────────────────────────────┐     ┌──────────────────┐

package/docs/anthropic-third-party-harness-changes.md

@@ -31,7 +31,7 @@ This is exactly the problem [ClawRouter](https://github.com/BlockRunAI/ClawRoute
 
 ## ClawRouter: Smart Routing for Agents
 
-ClawRouter is an open-source local proxy that sits between your agent and <!-- br:models.chatVisible -->73<!-- /br:models.chatVisible --> LLM models across 9 providers. It analyzes every request across 15 dimensions and routes it to the cheapest model that can handle it — in under 1ms, entirely locally.
+ClawRouter is an open-source local proxy that sits between your agent and <!-- br:models.chatVisible -->76<!-- /br:models.chatVisible --> LLM models across 9 providers. It analyzes every request across 15 dimensions and routes it to the cheapest model that can handle it — in under 1ms, entirely locally.
 
 ```
 Your Agent → ClawRouter (localhost:8402) → Best model for the job
@@ -56,7 +56,7 @@ From 20,000+ production requests:
 | Free models (trivial tasks)      | 12.8%        | $0.00             |
 | Others                           | 13.8%        | varies            |
 
-**Result: <!-- br:savings.autoVsBaselinePct -->88<!-- /br:savings.autoVsBaselinePct -->% cheaper than pinning Claude Opus 5 for every request** — the published figure, priced on a stated workload mix rather than estimated.
+**Result: <!-- br:savings.autoVsBaselinePct -->84<!-- /br:savings.autoVsBaselinePct -->% cheaper than pinning Claude Opus 5 for every request** — the published figure, priced on a stated workload mix rather than estimated.
 
 A typical user running 10K mixed requests/month:
 
@@ -74,7 +74,7 @@ Then point your agent at `http://localhost:8402/v1/` with any OpenAI-compatible
 
 - **No API keys to manage** — wallet-based cryptographic auth
 - **No subscriptions** — pay per request in USDC (Base or Solana)
-- **No vendor lock-in** — <!-- br:models.chatVisible -->73<!-- /br:models.chatVisible --> models, switch anytime
+- **No vendor lock-in** — <!-- br:models.chatVisible -->76<!-- /br:models.chatVisible --> models, switch anytime
 - **You control your wallet** — non-custodial, funds never held by a third party
 - **Budget caps** — set a max spend per session, ClawRouter gracefully downgrades when budget runs low
 - **Token compression** — 7-layer pipeline reduces token costs by 7–40% before they hit any provider

package/docs/clawrouter-cuts-llm-api-costs-500x.md

@@ -135,7 +135,7 @@ Less context repeated = fewer tokens = lower cost. Agents that need to recall ea
 
 ### 7. x402 Micropayments — Wallet-Based Budget Control
 
-<p align="center"><img src="assets/blockrun-clawrouter-x402-usdc-micropayment-wallet-budget-control.png" alt="Budget limits enforced by physical construction — wallet loaded via Base/Solana, pay per call across <!-- br:models.chatVisible -->73<!-- /br:models.chatVisible --> models, balance hits zero and the valve shuts cleanly" width="720"></p>
+<p align="center"><img src="assets/blockrun-clawrouter-x402-usdc-micropayment-wallet-budget-control.png" alt="Budget limits enforced by physical construction — wallet loaded via Base/Solana, pay per call across <!-- br:models.chatVisible -->76<!-- /br:models.chatVisible --> models, balance hits zero and the valve shuts cleanly" width="720"></p>
 
 ClawRouter pays for inference via [x402](https://x402.org/) USDC micropayments (Base or Solana). You load a wallet. Each inference call costs exactly what it costs. When the wallet runs low, requests stop cleanly.
 

package/docs/clawrouter-vs-openrouter-llm-routing-comparison.md

@@ -222,7 +222,7 @@ OpenRouter doesn't always pass through provider-specific features correctly. Ima
 
 **Direct provider routing.** ClawRouter routes through BlockRun's API directly to providers — not through a second aggregator. One hop, not two. Provider-specific features work because there's no middleman translating them.
 
-![Guaranteed Feature Parity & Direct Connectivity — Three-panel diagram: Vision (image_url auto-detected → vision-capable models only), Tool Calling (toolCalling flag → agentic models only), Catalog (curated <!-- br:models.chatVisible -->73<!-- /br:models.chatVisible --> models with automatic legacy-to-modern redirects). Direct provider routing means no dropped payloads.](./assets/clawrouter-feature-parity-direct-connectivity.png)
+![Guaranteed Feature Parity & Direct Connectivity — Three-panel diagram: Vision (image_url auto-detected → vision-capable models only), Tool Calling (toolCalling flag → agentic models only), Catalog (curated <!-- br:models.chatVisible -->76<!-- /br:models.chatVisible --> models with automatic legacy-to-modern redirects). Direct provider routing means no dropped payloads.](./assets/clawrouter-feature-parity-direct-connectivity.png)
 
 ---
 
@@ -240,7 +240,7 @@ When new models launch, OpenRouter's catalog lags. Users configure a model that
 
 ### How ClawRouter Solves This
 
-ClawRouter maintains a curated catalog of <!-- br:models.chatVisible -->73<!-- /br:models.chatVisible --> models across 9 providers (including <!-- br:models.free -->7<!-- /br:models.free --> free models), updated with each release. Delisted models have automatic redirect aliases:
+ClawRouter maintains a curated catalog of <!-- br:models.chatVisible -->76<!-- /br:models.chatVisible --> models across 9 providers (including <!-- br:models.free -->7<!-- /br:models.free --> free models), updated with each release. Delisted models have automatic redirect aliases:
 
 ```typescript
 // Delisted models redirect automatically
@@ -268,10 +268,10 @@ No silent drops. No stale catalog. Models are benchmarked for speed, quality, an
 | **Rate limits**     | Per-key, shared                  | Per-wallet, independent                                                                                  |
 | **Vision support**  | Images sometimes dropped         | Auto-detected, vision-only fallback                                                                      |
 | **Tool calling**    | Silent failures with some models | Flag-based filtering, guaranteed support                                                                 |
-| **Model catalog**   | Laggy, silent drops              | Curated <!-- br:models.chatVisible -->73<!-- /br:models.chatVisible --> models, redirect aliases         |
+| **Model catalog**   | Laggy, silent drops              | Curated <!-- br:models.chatVisible -->76<!-- /br:models.chatVisible --> models, redirect aliases         |
 | **Budget control**  | Monthly invoice                  | Per-session cap (`maxCostPerRun`)                                                                        |
 | **Setup**           | Create account, paste key        | Agent generates wallet, auto-configured                                                                  |
-| **Average cost**    | $25/M tokens (Opus direct)       | auto-routed = **<!-- br:savings.autoVsBaselinePct -->88<!-- /br:savings.autoVsBaselinePct -->% savings** |
+| **Average cost**    | $25/M tokens (Opus direct)       | auto-routed = **<!-- br:savings.autoVsBaselinePct -->84<!-- /br:savings.autoVsBaselinePct -->% savings** |
 
 ![The Engineering Matrix — Side-by-side feature comparison: OpenRouter vs ClawRouter across Routing, Authentication, Payment, Fallback, Model IDs, Empty Wallet, Vision/Tools, and Average Cost. ClawRouter wins on every dimension.](./assets/clawrouter-engineering-matrix-comparison.png)
 

package/docs/routing-profiles.md

@@ -75,11 +75,13 @@ Set `routing.overrides.agenticMode: false` to disable the agentic tier set, or `
 
 ## FREE (`/model free`)
 
-`free` is an alias, not a routed profile: it pins the free-tier default, currently **`free/nemotron-3.5-lightning`** — the same model that opens ECO SIMPLE. If that model is excluded (`/exclude add lightning`) or the budget cap forces a free fallback, the proxy walks the free cascade in this order: nemotron-3.5-lightning → 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). All seven are $0.00 and need no USDC.
+`free` is an alias, not a routed profile: it pins the free-tier default, currently **`free/nemotron-3.5-lightning`** — the same model that opens ECO SIMPLE. If that model is excluded (`/exclude add lightning`) or the budget cap forces a free fallback, the proxy walks the free cascade in this order: nemotron-3.5-lightning → nemotron-3-nano-30b → laguna-xs-2.1 → north-mini-code → nemotron-3-nano-omni-30b-a3b-reasoning → nemotron-3-ultra-550b → llama-3.2-11b-vision. All seven are $0.00 and need no USDC, and all seven are treated as **text-only** — see the vision note below.
 
 The order is not arbitrary. The head is whatever the gateway itself redirects the previous head to, so the proxy and the gateway never name different models. After it comes the fastest rung, then the two sub-second coders — adjacent on purpose, because they sit on _different_ capacity pools (our own NVIDIA key vs OpenRouter's $0 pool), so one pool's outage cannot take both. The vision model sits mid-chain, and the two slowest rungs go last. The tier stopped being NVIDIA-only on 2026-08-30: two of the seven come from Cohere and Poolside.
 
-Pin any of them directly with `/model lightning`, `/model nano-30b`, `/model laguna`, `/model north-mini`, `/model nemotron-omni` (vision), `/model ultra-550b` or `/model llama-vision` (vision).
+Pin any of them directly with `/model lightning`, `/model nano-30b`, `/model laguna`, `/model north-mini`, `/model nemotron-omni`, `/model ultra-550b` or `/model llama-vision`.
+
+**No free model carries the `vision` flag,** even the two the upstream catalogs advertise as vision-capable. A 64×64 solid-colour probe on 2026-08-31 got the right answer 1 time in 4 from `nemotron-3-nano-omni` on Base, and "white" for red on Solana — where the response's own `model` field showed a silent fallback to a text model. `llama-3.2-11b-vision`, despite the name, said "I'm unable to see the image" on 3 of 3 while a plain-text control passed. Every one of those is an HTTP 200, so the caller gets a confident wrong answer with nothing to branch on. Since `vision` is what `filterByVision()` uses to _aim_ image turns at a model, flagging them would route real traffic into that. Requests carrying an `image_url` go to a paid vision model.
 
 ---
 
@@ -111,7 +113,7 @@ Combined input + output rate per 1M tokens, primaries only:
 | COMPLEX   | $1.75 | $14.00 | **88%**  |
 | REASONING | $0.70 | $0.70  | 0%       |
 
-The published savings claim (**88%** on `auto`, **98%** on `eco`, against pinning Claude Opus 5) is blended across a stated workload mix and priced on the live catalog — see [`savings-mix.json`](https://github.com/BlockRunAI/blockrun/blob/main/src/brand/savings-mix.json). Per-tier numbers above are illustrative; the blended figure is the one to quote.
+The published savings claim (**84%** on `auto`, **98%** on `eco`, against pinning Claude Opus 5) is blended across a stated workload mix and priced on the live catalog — see [`savings-mix.json`](https://github.com/BlockRunAI/blockrun/blob/main/src/brand/savings-mix.json). Per-tier numbers above are illustrative; the blended figure is the one to quote.
 
 ---
 

package/package.json

@@ -1,7 +1,7 @@
 {
   "name": "@blockrun/clawrouter",
-  "version": "0.12.258",
-  "description": "Smart LLM router — save 88% on inference costs. 73 models (7 free), one wallet, x402 micropayments.",
+  "version": "0.12.260",
+  "description": "Smart LLM router — save 84% on inference costs. 76 models (7 free), one wallet, x402 micropayments.",
   "type": "module",
   "main": "dist/index.js",
   "types": "dist/index.d.ts",

package/skills/clawrouter/SKILL.md

@@ -1,6 +1,6 @@
 ---
 name: clawrouter
-description: Hosted-gateway LLM router — save 88% on inference costs. A local proxy that forwards each request to the blockrun.ai gateway, which routes to the cheapest capable model across 73 models from OpenAI, Anthropic, Google, DeepSeek, xAI, Z.AI, and more. 7 free open-weight models included. Also exposes realtime market data (global stocks, crypto, FX, commodities), Twitter/X intelligence, prediction-market data across Polymarket, Kalshi, Limitless, Opinion, Predict.Fun, dFlow + UMA oracle resolution + wallet identity & clustering, phone-number intelligence (carrier + SIM-swap fraud detection) plus AI-powered outbound voice calls (Twilio + Bland.ai), AND the Surf unified crypto data API (84 endpoints — CEX/DEX, on-chain SQL over 80+ ClickHouse tables, 100M+ labeled wallets, prediction markets, social/CT mindshare, news, VC fund intel) as built-in agent tools. Not a local-inference tool — prompts are sent to the blockrun.ai gateway.
+description: Hosted-gateway LLM router — save 84% on inference costs. A local proxy that forwards each request to the blockrun.ai gateway, which routes to the cheapest capable model across 76 models from OpenAI, Anthropic, Google, DeepSeek, xAI, Z.AI, and more. 7 free open-weight models included. Also exposes realtime market data (global stocks, crypto, FX, commodities), Twitter/X intelligence, prediction-market data across Polymarket, Kalshi, Limitless, Opinion, Predict.Fun, dFlow + UMA oracle resolution + wallet identity & clustering, phone-number intelligence (carrier + SIM-swap fraud detection) plus AI-powered outbound voice calls (Twilio + Bland.ai), AND the Surf unified crypto data API (84 endpoints — CEX/DEX, on-chain SQL over 80+ ClickHouse tables, 100M+ labeled wallets, prediction markets, social/CT mindshare, news, VC fund intel) as built-in agent tools. Not a local-inference tool — prompts are sent to the blockrun.ai gateway.
 triggers:
   - "clawrouter"
   - "claw router"
@@ -53,7 +53,7 @@ metadata:
 
 # ClawRouter
 
-Hosted-gateway LLM router that saves <!-- br:savings.autoVsBaselinePct -->88<!-- /br:savings.autoVsBaselinePct -->% on inference costs by forwarding each request to the blockrun.ai gateway, which picks the cheapest model capable of handling it across <!-- br:models.chatVisible -->73<!-- /br:models.chatVisible --> models from 9 providers (<!-- br:models.free -->7<!-- /br:models.free --> free open-weight models). All billing flows through one USDC wallet; you do not hold provider API keys.
+Hosted-gateway LLM router that saves <!-- br:savings.autoVsBaselinePct -->84<!-- /br:savings.autoVsBaselinePct -->% on inference costs by forwarding each request to the blockrun.ai gateway, which picks the cheapest model capable of handling it across <!-- br:models.chatVisible -->76<!-- /br:models.chatVisible --> models from 9 providers (<!-- br:models.free -->7<!-- /br:models.free --> free open-weight models). All billing flows through one USDC wallet; you do not hold provider API keys.
 
 **This is not a local-inference tool.** ClawRouter is a thin local proxy. Your prompts are sent over HTTPS to the blockrun.ai gateway for model execution. If your workload requires inference that never leaves your machine, use a local runtime like Ollama — ClawRouter is not the right tool for that use case.
 
@@ -147,7 +147,7 @@ it against a baseline nobody wrote down. See
 
 ## Available Models
 
-<!-- br:models.chatVisible -->73<!-- /br:models.chatVisible --> models including: gpt-5.6-terra [balanced, stable default], gpt-5.6-sol [flagship], gpt-5.6-luna [cost-efficient], gpt-5.6-sol/terra/luna-pro [pro reasoning tiers], gpt-5.5, gpt-5.5-pro [max compute], chat-latest [ChatGPT Instant], gpt-5.4, gpt-4o, o3, claude-fable-5, claude-opus-5 [Anthropic flagship, 1M ctx], claude-opus-4.8, claude-opus-4.7, claude-opus-4.5, claude-sonnet-5, claude-sonnet-4.6, gemini-3.1-pro, gemini-3.6-flash, gemini-3.5-flash-lite, gemini-2.5-flash, deepseek-v4-pro, deepseek-chat, grok-4.5, grok-4.3, grok-build-0.1, glm-5.3 [Z.AI flagship, 1M ctx], glm-5.3-flash [natively multimodal, 1M ctx — $0.15/$0.50], glm-5.2, minimax-m3, kimi-k3 [1M ctx flagship], kimi-k2.7, qwen3.7-max [Qwen flagship, 1M ctx], qwen3.7-plus, qwen3.7-flash, hy3 [Tencent], mimo [Xiaomi MiMo-V2.5 Pro, 1M ctx], and 7 free open-weight models (nemotron-3.5-lightning [1M ctx, thinking mode — the free default], nemotron-3-nano-30b [fastest, ~121 tok/s], nemotron-3-ultra-550b [550B/55B MoE, 1M ctx — the largest free model], nemotron-3-nano-omni-30b-a3b-reasoning [256K ctx, vision — text + images + video + audio], llama-3.2-11b-vision [Meta Llama, image input], north-mini-code [Cohere, coding, sub-second], laguna-xs-2.1 [Poolside, coding, ~161 tok/s]).
+<!-- br:models.chatVisible -->76<!-- /br:models.chatVisible --> models including: gpt-5.6-terra [balanced, stable default], gpt-5.6-sol [flagship], gpt-5.6-luna [cost-efficient], gpt-5.6-sol/terra/luna-pro [pro reasoning tiers], gpt-5.5, gpt-5.5-pro [max compute], chat-latest [ChatGPT Instant], gpt-5.4, gpt-4o, o3, claude-fable-5, claude-opus-5 [Anthropic flagship, 1M ctx], claude-opus-4.8, claude-opus-4.7, claude-opus-4.5, claude-sonnet-5, claude-sonnet-4.6, gemini-3.1-pro, gemini-3.6-flash, gemini-3.5-flash-lite, gemini-2.5-flash, deepseek-v4-pro, deepseek-chat, grok-4.5, grok-4.3, grok-build-0.1, glm-5.3 [Z.AI flagship, 1M ctx], glm-5.3-flash [natively multimodal, 1M ctx — $0.15/$0.50], glm-5.2, minimax-m3, kimi-k3 [1M ctx flagship], kimi-k2.7, qwen3.7-max [Qwen flagship, 1M ctx], qwen3.7-plus, qwen3.7-flash, hy3 [Tencent], mimo [Xiaomi MiMo-V2.5 Pro, 1M ctx], and 7 free open-weight models (nemotron-3.5-lightning [1M ctx, thinking mode — the free default], nemotron-3-nano-30b [fastest, ~121 tok/s], nemotron-3-ultra-550b [550B/55B MoE, 1M ctx — the largest free model], nemotron-3-nano-omni-30b-a3b-reasoning [256K ctx], llama-3.2-11b-vision [Meta Llama], north-mini-code [Cohere, coding, sub-second], laguna-xs-2.1 [Poolside, coding, ~161 tok/s]). The free tier is text-only — image turns route to a paid vision model, because the two free models the catalogs advertise as vision-capable fail a solid-colour probe on both chains and do so with an HTTP 200.
 
 ## Built-in Agent Tools
 

package/src/top-models.json

@@ -33,10 +33,13 @@
   "zai/glm-5.1",
   "zai/glm-5-turbo",
   "zai/glm-5",
+  "xiaomi/mimo-v2.5",
   "minimax/minimax-m3",
   "minimax/minimax-m2.7",
   "moonshot/kimi-k3",
   "qwen/qwen3.7-max",
+  "qwen/qwen3.8-flash",
+  "deepseek/deepseek-v4-flash-vision-exp",
   "deepseek/deepseek-v4-pro",
   "deepseek/deepseek-chat",
   "deepseek/deepseek-reasoner",