@blockrun/llm 3.13.5 → 3.13.6

package/README.md

@@ -218,7 +218,7 @@ console.log(`Saved ${(result.routing.savings * 100).toFixed(0)}%`); // this requ
 const complex = await client.smartChat('Prove the Riemann hypothesis step by step');
 console.log(complex.model);  // 'xai/grok-4.3'
 
-// Inspect how the request was classified and ranked (Router v3.4 portfolio).
+// Inspect how the request was classified and ranked (Router Core V3.5 portfolio).
 console.log(complex.routing.method);     // 'portfolio'
 console.log(complex.routing.taskType);   // 'reasoning'
 console.log(complex.routing.candidates); // ranked, capability-eligible models
@@ -279,8 +279,8 @@ flowchart LR
     F --> G["response<br/>+ full routing metadata"]
 ```
 
-Since ClawRouter v0.12.242, Auto uses the deterministic **Router v3.4 portfolio
-strategy**: it classifies the task shape locally across
+Auto uses the deterministic **Router Core V3.5 portfolio strategy**
+(`DEFAULT_ROUTING_CONFIG.version` in the pinned engine): it classifies the task shape locally across
 <!-- br:clawrouter.dimensions -->15<!-- /br:clawrouter.dimensions --> dimensions
 (token count, code presence, reasoning markers, technical/creative terms,
 agentic patterns, …), enforces tool / vision / structured-output / context
@@ -296,7 +296,7 @@ anchors the portfolio's candidate pool):
 
 | Tier | Example Tasks | ECO | AUTO | PREMIUM |
 |------|---------------|-----|------|---------|
-| SIMPLE | "What is 2+2?", definitions | step-3.7-flash (**FREE**) | gemini-2.5-flash ($0.30/$2.50) | gemini-3.5-flash ($1.50/$9) |
+| SIMPLE | "What is 2+2?", definitions | nemotron-3.5-lightning (**FREE**) | gemini-2.5-flash ($0.30/$2.50) | gemini-3.5-flash ($1.50/$9) |
 | MEDIUM | Code snippets, explanations | glm-5.3-flash ($0.15/$0.50) | gemini-3.5-flash ($1.50/$9) | gpt-5.3-codex ($1.75/$14.00) |
 | COMPLEX | Architecture, long documents | glm-5.3-flash ($0.15/$0.50) | gemini-3.1-pro ($2/$12) | claude-fable-5 ($10/$50) |
 | REASONING | Proofs, multi-step reasoning | deepseek-reasoner ($0.14/$0.28) | deepseek-reasoner ($0.14/$0.28) | claude-sonnet-5 ($3/$15) |

package/dist/index.cjs

@@ -145,7 +145,7 @@ var APIError = class extends BlockrunError {
   }
 };
 
-// node_modules/.pnpm/@blockrun+router-core@https+++codeload.github.com+BlockRunAI+router-core+tar.gz+5d911879d7f1e_fbldkf3f3jmlwtwu53ftq2mj24/node_modules/@blockrun/router-core/dist/index.js
+// node_modules/.pnpm/@blockrun+router-core@https+++codeload.github.com+BlockRunAI+router-core+tar.gz+814fd3b6bc788_xq5fkcckbb2p7dtp6dodey4yri/node_modules/@blockrun/router-core/dist/index.js
 function scoreTokenCount(estimatedTokens, thresholds) {
   if (estimatedTokens < thresholds.simple) {
     return { name: "tokenCount", score: -1, signal: `short (${estimatedTokens} tokens)` };
@@ -668,6 +668,13 @@ var DEFAULT_MODEL_CAPABILITIES = Object.freeze({
     supportsTools: true,
     supportsVision: true
   },
+  "cohere/north-mini-code": {
+    // supportsTools: not probed — fails closed
+    contextWindow: 256e3,
+    maxOutputTokens: 16384,
+    supportsTools: false,
+    supportsVision: false
+  },
   "deepseek/deepseek-chat": {
     contextWindow: 1048576,
     maxOutputTokens: 65536,
@@ -758,8 +765,15 @@ var DEFAULT_MODEL_CAPABILITIES = Object.freeze({
     supportsTools: true,
     supportsVision: true
   },
-  "nvidia/mistral-nemotron": {
-    // supportsTools: gateway unavailable at probe time — fails closed
+  "nvidia/llama-3.2-11b-vision": {
+    // supportsTools: not probed — fails closed
+    contextWindow: 128e3,
+    maxOutputTokens: 16384,
+    supportsTools: false,
+    supportsVision: true
+  },
+  "nvidia/nemotron-3-nano-30b": {
+    // supportsTools: not probed — fails closed
     contextWindow: 131072,
     maxOutputTokens: 16384,
     supportsTools: false,
@@ -771,21 +785,16 @@ var DEFAULT_MODEL_CAPABILITIES = Object.freeze({
     supportsTools: false,
     supportsVision: true
   },
-  "nvidia/nemotron-nano-12b-v2-vl": {
-    // supportsTools: gateway unavailable at probe time — fails closed
-    contextWindow: 131072,
-    maxOutputTokens: 16384,
-    supportsTools: false,
-    supportsVision: true
-  },
-  "nvidia/nemotron-nano-9b-v2": {
-    contextWindow: 131072,
+  "nvidia/nemotron-3-ultra-550b": {
+    // supportsTools: not probed — fails closed
+    contextWindow: 1e6,
     maxOutputTokens: 16384,
     supportsTools: false,
     supportsVision: false
   },
-  "nvidia/step-3.7-flash": {
-    contextWindow: 131072,
+  "nvidia/nemotron-3.5-lightning": {
+    // supportsTools: not probed — fails closed
+    contextWindow: 1e6,
     maxOutputTokens: 16384,
     supportsTools: false,
     supportsVision: false
@@ -845,13 +854,6 @@ var DEFAULT_MODEL_CAPABILITIES = Object.freeze({
     supportsTools: false,
     supportsVision: true
   },
-  "openai/gpt-5.3": {
-    // supportsTools: gateway unavailable at probe time — fails closed
-    contextWindow: 128e3,
-    maxOutputTokens: 128e3,
-    supportsTools: false,
-    supportsVision: true
-  },
   "openai/gpt-5.3-codex": {
     // supportsTools: gateway unavailable at probe time — fails closed; override: 2026-08-29 probe: every request (6 plain + 3 tool attempts) returned a gateway 500, so the probe measured an incident, not the model. Codex's function calling is established by the 2026-07 Terminal-Bench / tau2 calibration trajectories in portfolio.ts. Hosts observing the 500s should drop it with RouterOptions.unavailableModels rather than this snapshot claiming the model cannot call tools.
     contextWindow: 4e5,
@@ -957,6 +959,13 @@ var DEFAULT_MODEL_CAPABILITIES = Object.freeze({
     supportsTools: true,
     supportsVision: false
   },
+  "poolside/laguna-xs-2.1": {
+    // supportsTools: not probed — fails closed
+    contextWindow: 131072,
+    maxOutputTokens: 16384,
+    supportsTools: false,
+    supportsVision: false
+  },
   "qwen/qwen3.7-flash": {
     contextWindow: 1e6,
     maxOutputTokens: 65536,
@@ -3498,8 +3507,8 @@ var DEFAULT_ROUTING_CONFIG = {
         // $0.20/$1.25, 1M ctx
         "google/gemini-2.5-flash-lite",
         // $0.10/$0.40
-        "nvidia/step-3.7-flash"
-        // FREE backstop — NVIDIA free tier (probed 2026-08-21)
+        "nvidia/nemotron-3.5-lightning"
+        // FREE backstop — NVIDIA free tier (probed 2026-08-30)
       ]
     },
     MEDIUM: {
@@ -3588,15 +3597,19 @@ var DEFAULT_ROUTING_CONFIG = {
   // Eco tier configs - absolute cheapest (blockrun/eco)
   ecoTiers: {
     SIMPLE: {
-      primary: "nvidia/step-3.7-flash",
-      // FREE — NVIDIA free tier flagship
+      primary: "nvidia/nemotron-3.5-lightning",
+      // FREE — NVIDIA free tier flagship, 1M ctx
       fallback: [
-        "nvidia/nemotron-nano-9b-v2",
-        // FREE — compact + fast, high-volume light tasks
+        "nvidia/nemotron-3-nano-30b",
+        // FREE — fastest free model (~121 tok/s)
         // The free head keeps rotting with NVIDIA's hosting (deepseek-v4-flash
         // 410 2026-08-12, seed-oss-36b 410 2026-08-03, gpt-oss-120b/20b 400
-        // 2026-08-21). Each retirement retargets the two free rungs to the
-        // current free tier; the paid rungs below never move.
+        // 2026-08-21, and on 2026-08-30 FOUR of the five visible free models at
+        // once — step-3.7-flash, nemotron-nano-9b-v2 and nemotron-nano-12b-v2-vl
+        // all 410, mistral-nemotron hung). Each retirement retargets the two
+        // free rungs to the current free tier; the paid rungs below never move.
+        // The head follows blockrun's own redirect of the model it replaces, so
+        // the router and the gateway never name different models.
         "google/gemini-2.5-flash-lite",
         // $0.10/$0.40 — cheapest paid rung
         "zai/glm-5.3-flash",
@@ -3740,7 +3753,7 @@ var DEFAULT_ROUTING_CONFIG = {
         // strongest open-weight reasoner
         "deepseek/deepseek-chat",
         // Cheap, reliable
-        "nvidia/step-3.7-flash"
+        "nvidia/nemotron-3.5-lightning"
         // NVIDIA free ultimate backstop
       ]
     },
@@ -3844,7 +3857,7 @@ var DEFAULT_ROUTING_CONFIG = {
         // retail high-risk 3/3
         "deepseek/deepseek-chat",
         // cheap, reliable
-        "nvidia/step-3.7-flash"
+        "nvidia/nemotron-3.5-lightning"
         // NVIDIA free ultimate backstop
       ]
     },
@@ -4465,7 +4478,7 @@ function getCostSummary() {
 }
 
 // src/version.ts
-var SDK_VERSION = "3.13.5";
+var SDK_VERSION = "3.13.6";
 var USER_AGENT = `blockrun-ts/${SDK_VERSION}`;
 
 // src/client.ts

package/dist/index.js

@@ -31,7 +31,7 @@ var APIError = class extends BlockrunError {
   }
 };
 
-// node_modules/.pnpm/@blockrun+router-core@https+++codeload.github.com+BlockRunAI+router-core+tar.gz+5d911879d7f1e_fbldkf3f3jmlwtwu53ftq2mj24/node_modules/@blockrun/router-core/dist/index.js
+// node_modules/.pnpm/@blockrun+router-core@https+++codeload.github.com+BlockRunAI+router-core+tar.gz+814fd3b6bc788_xq5fkcckbb2p7dtp6dodey4yri/node_modules/@blockrun/router-core/dist/index.js
 function scoreTokenCount(estimatedTokens, thresholds) {
   if (estimatedTokens < thresholds.simple) {
     return { name: "tokenCount", score: -1, signal: `short (${estimatedTokens} tokens)` };
@@ -554,6 +554,13 @@ var DEFAULT_MODEL_CAPABILITIES = Object.freeze({
     supportsTools: true,
     supportsVision: true
   },
+  "cohere/north-mini-code": {
+    // supportsTools: not probed — fails closed
+    contextWindow: 256e3,
+    maxOutputTokens: 16384,
+    supportsTools: false,
+    supportsVision: false
+  },
   "deepseek/deepseek-chat": {
     contextWindow: 1048576,
     maxOutputTokens: 65536,
@@ -644,8 +651,15 @@ var DEFAULT_MODEL_CAPABILITIES = Object.freeze({
     supportsTools: true,
     supportsVision: true
   },
-  "nvidia/mistral-nemotron": {
-    // supportsTools: gateway unavailable at probe time — fails closed
+  "nvidia/llama-3.2-11b-vision": {
+    // supportsTools: not probed — fails closed
+    contextWindow: 128e3,
+    maxOutputTokens: 16384,
+    supportsTools: false,
+    supportsVision: true
+  },
+  "nvidia/nemotron-3-nano-30b": {
+    // supportsTools: not probed — fails closed
     contextWindow: 131072,
     maxOutputTokens: 16384,
     supportsTools: false,
@@ -657,21 +671,16 @@ var DEFAULT_MODEL_CAPABILITIES = Object.freeze({
     supportsTools: false,
     supportsVision: true
   },
-  "nvidia/nemotron-nano-12b-v2-vl": {
-    // supportsTools: gateway unavailable at probe time — fails closed
-    contextWindow: 131072,
-    maxOutputTokens: 16384,
-    supportsTools: false,
-    supportsVision: true
-  },
-  "nvidia/nemotron-nano-9b-v2": {
-    contextWindow: 131072,
+  "nvidia/nemotron-3-ultra-550b": {
+    // supportsTools: not probed — fails closed
+    contextWindow: 1e6,
     maxOutputTokens: 16384,
     supportsTools: false,
     supportsVision: false
   },
-  "nvidia/step-3.7-flash": {
-    contextWindow: 131072,
+  "nvidia/nemotron-3.5-lightning": {
+    // supportsTools: not probed — fails closed
+    contextWindow: 1e6,
     maxOutputTokens: 16384,
     supportsTools: false,
     supportsVision: false
@@ -731,13 +740,6 @@ var DEFAULT_MODEL_CAPABILITIES = Object.freeze({
     supportsTools: false,
     supportsVision: true
   },
-  "openai/gpt-5.3": {
-    // supportsTools: gateway unavailable at probe time — fails closed
-    contextWindow: 128e3,
-    maxOutputTokens: 128e3,
-    supportsTools: false,
-    supportsVision: true
-  },
   "openai/gpt-5.3-codex": {
     // supportsTools: gateway unavailable at probe time — fails closed; override: 2026-08-29 probe: every request (6 plain + 3 tool attempts) returned a gateway 500, so the probe measured an incident, not the model. Codex's function calling is established by the 2026-07 Terminal-Bench / tau2 calibration trajectories in portfolio.ts. Hosts observing the 500s should drop it with RouterOptions.unavailableModels rather than this snapshot claiming the model cannot call tools.
     contextWindow: 4e5,
@@ -843,6 +845,13 @@ var DEFAULT_MODEL_CAPABILITIES = Object.freeze({
     supportsTools: true,
     supportsVision: false
   },
+  "poolside/laguna-xs-2.1": {
+    // supportsTools: not probed — fails closed
+    contextWindow: 131072,
+    maxOutputTokens: 16384,
+    supportsTools: false,
+    supportsVision: false
+  },
   "qwen/qwen3.7-flash": {
     contextWindow: 1e6,
     maxOutputTokens: 65536,
@@ -3384,8 +3393,8 @@ var DEFAULT_ROUTING_CONFIG = {
         // $0.20/$1.25, 1M ctx
         "google/gemini-2.5-flash-lite",
         // $0.10/$0.40
-        "nvidia/step-3.7-flash"
-        // FREE backstop — NVIDIA free tier (probed 2026-08-21)
+        "nvidia/nemotron-3.5-lightning"
+        // FREE backstop — NVIDIA free tier (probed 2026-08-30)
       ]
     },
     MEDIUM: {
@@ -3474,15 +3483,19 @@ var DEFAULT_ROUTING_CONFIG = {
   // Eco tier configs - absolute cheapest (blockrun/eco)
   ecoTiers: {
     SIMPLE: {
-      primary: "nvidia/step-3.7-flash",
-      // FREE — NVIDIA free tier flagship
+      primary: "nvidia/nemotron-3.5-lightning",
+      // FREE — NVIDIA free tier flagship, 1M ctx
       fallback: [
-        "nvidia/nemotron-nano-9b-v2",
-        // FREE — compact + fast, high-volume light tasks
+        "nvidia/nemotron-3-nano-30b",
+        // FREE — fastest free model (~121 tok/s)
         // The free head keeps rotting with NVIDIA's hosting (deepseek-v4-flash
         // 410 2026-08-12, seed-oss-36b 410 2026-08-03, gpt-oss-120b/20b 400
-        // 2026-08-21). Each retirement retargets the two free rungs to the
-        // current free tier; the paid rungs below never move.
+        // 2026-08-21, and on 2026-08-30 FOUR of the five visible free models at
+        // once — step-3.7-flash, nemotron-nano-9b-v2 and nemotron-nano-12b-v2-vl
+        // all 410, mistral-nemotron hung). Each retirement retargets the two
+        // free rungs to the current free tier; the paid rungs below never move.
+        // The head follows blockrun's own redirect of the model it replaces, so
+        // the router and the gateway never name different models.
         "google/gemini-2.5-flash-lite",
         // $0.10/$0.40 — cheapest paid rung
         "zai/glm-5.3-flash",
@@ -3626,7 +3639,7 @@ var DEFAULT_ROUTING_CONFIG = {
         // strongest open-weight reasoner
         "deepseek/deepseek-chat",
         // Cheap, reliable
-        "nvidia/step-3.7-flash"
+        "nvidia/nemotron-3.5-lightning"
         // NVIDIA free ultimate backstop
       ]
     },
@@ -3730,7 +3743,7 @@ var DEFAULT_ROUTING_CONFIG = {
         // retail high-risk 3/3
         "deepseek/deepseek-chat",
         // cheap, reliable
-        "nvidia/step-3.7-flash"
+        "nvidia/nemotron-3.5-lightning"
         // NVIDIA free ultimate backstop
       ]
     },
@@ -4351,7 +4364,7 @@ function getCostSummary() {
 }
 
 // src/version.ts
-var SDK_VERSION = "3.13.5";
+var SDK_VERSION = "3.13.6";
 var USER_AGENT = `blockrun-ts/${SDK_VERSION}`;
 
 // src/client.ts

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@blockrun/llm",
-  "version": "3.13.5",
+  "version": "3.13.6",
   "type": "module",
   "description": "BlockRun SDK - Pay-per-request AI (LLM, Image, Video, Music, Voice) via x402 on Base and Solana",
   "main": "dist/index.cjs",
@@ -56,7 +56,7 @@
   },
   "devDependencies": {
     "@blockrun/core": "^0.1.0",
-    "@blockrun/router-core": "https://codeload.github.com/BlockRunAI/router-core/tar.gz/5d911879d7f1eddf93d513b08798998320272cc8",
+    "@blockrun/router-core": "https://codeload.github.com/BlockRunAI/router-core/tar.gz/814fd3b6bc7880a84bbe030bd08c98e8670031e7",
     "@eslint/js": "^9.39.5",
     "@types/node": "^20.19.43",
     "eslint": "^9.39.5",