pi-ollama-cloud 0.12.0 → 0.12.2

package/CHANGELOG.md

@@ -2,6 +2,14 @@
 
 All notable changes to this project will be documented in this file.
 
+## [0.12.2] - 2026-09-27
+
+- Refresh the model catalog for the September 25, 2026 retirements (`deepseek-v4-flash:0731`, `glm-5.1`, `qwen3.5:397b`): dropped from the generated catalog and pricing, 20 -> 17 models. Recommended replacements per Ollama: `deepseek-v4.1-flash`, `glm-5.3`, and `glm-5.3-flash` / `deepseek-v4.1-flash`.
+
+## [0.12.1] - 2026-09-14
+
+- Omit empty `openRouterRouting` / `vercelGatewayRouting` from `buildCompat` (set to `undefined`, not `{}`): pi-ai reads the raw `model.compat` and treats `{}` as truthy, sending a stray `provider: {}` on every Ollama chat completion. Fixes #60. Thanks @0xbentang (#61).
+
 ## [0.12.0] - 2026-09-11
 
 - Source per-model thinking levels from models.dev instead of hardcoded maps. `scripts/generate-reasoning.ts` fetches the `ollama-cloud` provider's `reasoning_options` into `reasoning.generated.ts`, and `thinking-levels.ts` maps each model's effort values onto Pi's levels (toggle-only models become a binary on/off map; models with no models.dev entry fall back to `DEFAULT`). The `off` switch is handled by a small override table for models verified not to honor `reasoning_effort:"none"` (`gpt-oss:20b`, `gpt-oss:120b`, `minimax-m2.7`). Removed the now-stale per-family maps and the `docs/think-experiment.md` doc.

package/models.generated.ts

@@ -1,57 +1,16 @@
 // Auto-generated by scripts/generate-models.ts
 // Do not edit manually.
-// Generated: 2026-09-12T02:41:57.025Z
-// Model count: 20
+// Generated: 2026-09-27T00:52:07.773Z
+// Model count: 17
 
 import type { ProviderModelConfig } from "@earendil-works/pi-coding-agent";
 
 export const GENERATED_MODELS: ProviderModelConfig[] = [
-  {
-    id: "deepseek-v4-flash:0731",
-    name: "deepseek-v4-flash:0731",
-    compat: {
-      maxTokensField: "max_tokens",
-      openRouterRouting: {},
-      requiresAssistantAfterToolResult: false,
-      requiresReasoningContentOnAssistantMessages: false,
-      requiresThinkingAsText: false,
-      requiresToolResultName: false,
-      sendSessionAffinityHeaders: false,
-      supportsDeveloperRole: false,
-      supportsLongCacheRetention: false,
-      supportsReasoningEffort: true,
-      supportsStore: false,
-      supportsStrictMode: false,
-      supportsUsageInStreaming: true,
-      thinkingFormat: "openai",
-      vercelGatewayRouting: {},
-      zaiToolStream: false,
-    },
-    contextWindow: 1048576,
-    cost: {
-      cacheRead: 0.014,
-      cacheWrite: 0,
-      input: 0.44,
-      output: 1.32,
-    },
-    input: ["text"],
-    maxTokens: 65536,
-    reasoning: true,
-    thinkingLevelMap: {
-      high: "high",
-      low: null,
-      medium: null,
-      minimal: null,
-      off: "none",
-      xhigh: "max",
-    },
-  },
   {
     id: "deepseek-v4-pro:0813",
     name: "deepseek-v4-pro:0813",
     compat: {
       maxTokensField: "max_tokens",
-      openRouterRouting: {},
       requiresAssistantAfterToolResult: false,
       requiresReasoningContentOnAssistantMessages: false,
       requiresThinkingAsText: false,
@@ -64,7 +23,6 @@ export const GENERATED_MODELS: ProviderModelConfig[] = [
       supportsStrictMode: false,
       supportsUsageInStreaming: true,
       thinkingFormat: "openai",
-      vercelGatewayRouting: {},
       zaiToolStream: false,
     },
     contextWindow: 1048576,
@@ -91,7 +49,6 @@ export const GENERATED_MODELS: ProviderModelConfig[] = [
     name: "deepseek-v4.1-flash",
     compat: {
       maxTokensField: "max_tokens",
-      openRouterRouting: {},
       requiresAssistantAfterToolResult: false,
       requiresReasoningContentOnAssistantMessages: false,
       requiresThinkingAsText: false,
@@ -104,7 +61,6 @@ export const GENERATED_MODELS: ProviderModelConfig[] = [
       supportsStrictMode: false,
       supportsUsageInStreaming: true,
       thinkingFormat: "openai",
-      vercelGatewayRouting: {},
       zaiToolStream: false,
     },
     contextWindow: 1048576,
@@ -131,7 +87,6 @@ export const GENERATED_MODELS: ProviderModelConfig[] = [
     name: "gemma4:31b",
     compat: {
       maxTokensField: "max_tokens",
-      openRouterRouting: {},
       requiresAssistantAfterToolResult: false,
       requiresReasoningContentOnAssistantMessages: false,
       requiresThinkingAsText: false,
@@ -144,7 +99,6 @@ export const GENERATED_MODELS: ProviderModelConfig[] = [
       supportsStrictMode: false,
       supportsUsageInStreaming: true,
       thinkingFormat: "openai",
-      vercelGatewayRouting: {},
       zaiToolStream: false,
     },
     contextWindow: 262144,
@@ -166,52 +120,11 @@ export const GENERATED_MODELS: ProviderModelConfig[] = [
       xhigh: null,
     },
   },
-  {
-    id: "glm-5.1",
-    name: "glm-5.1",
-    compat: {
-      maxTokensField: "max_tokens",
-      openRouterRouting: {},
-      requiresAssistantAfterToolResult: false,
-      requiresReasoningContentOnAssistantMessages: false,
-      requiresThinkingAsText: false,
-      requiresToolResultName: false,
-      sendSessionAffinityHeaders: false,
-      supportsDeveloperRole: false,
-      supportsLongCacheRetention: false,
-      supportsReasoningEffort: true,
-      supportsStore: false,
-      supportsStrictMode: false,
-      supportsUsageInStreaming: true,
-      thinkingFormat: "openai",
-      vercelGatewayRouting: {},
-      zaiToolStream: false,
-    },
-    contextWindow: 202752,
-    cost: {
-      cacheRead: 0.2,
-      cacheWrite: 0,
-      input: 1,
-      output: 3.2,
-    },
-    input: ["text"],
-    maxTokens: 131072,
-    reasoning: true,
-    thinkingLevelMap: {
-      high: null,
-      low: null,
-      medium: "medium",
-      minimal: null,
-      off: "none",
-      xhigh: null,
-    },
-  },
   {
     id: "glm-5.2",
     name: "glm-5.2",
     compat: {
       maxTokensField: "max_tokens",
-      openRouterRouting: {},
       requiresAssistantAfterToolResult: false,
       requiresReasoningContentOnAssistantMessages: false,
       requiresThinkingAsText: false,
@@ -224,7 +137,6 @@ export const GENERATED_MODELS: ProviderModelConfig[] = [
       supportsStrictMode: false,
       supportsUsageInStreaming: true,
       thinkingFormat: "openai",
-      vercelGatewayRouting: {},
       zaiToolStream: false,
     },
     contextWindow: 1048576,
@@ -251,7 +163,6 @@ export const GENERATED_MODELS: ProviderModelConfig[] = [
     name: "glm-5.3",
     compat: {
       maxTokensField: "max_tokens",
-      openRouterRouting: {},
       requiresAssistantAfterToolResult: false,
       requiresReasoningContentOnAssistantMessages: false,
       requiresThinkingAsText: false,
@@ -264,7 +175,6 @@ export const GENERATED_MODELS: ProviderModelConfig[] = [
       supportsStrictMode: false,
       supportsUsageInStreaming: true,
       thinkingFormat: "openai",
-      vercelGatewayRouting: {},
       zaiToolStream: false,
     },
     contextWindow: 1048576,
@@ -291,7 +201,6 @@ export const GENERATED_MODELS: ProviderModelConfig[] = [
     name: "glm-5.3-flash",
     compat: {
       maxTokensField: "max_tokens",
-      openRouterRouting: {},
       requiresAssistantAfterToolResult: false,
       requiresReasoningContentOnAssistantMessages: false,
       requiresThinkingAsText: false,
@@ -304,7 +213,6 @@ export const GENERATED_MODELS: ProviderModelConfig[] = [
       supportsStrictMode: false,
       supportsUsageInStreaming: true,
       thinkingFormat: "openai",
-      vercelGatewayRouting: {},
       zaiToolStream: false,
     },
     contextWindow: 1048576,
@@ -331,7 +239,6 @@ export const GENERATED_MODELS: ProviderModelConfig[] = [
     name: "gpt-oss:120b",
     compat: {
       maxTokensField: "max_tokens",
-      openRouterRouting: {},
       requiresAssistantAfterToolResult: false,
       requiresReasoningContentOnAssistantMessages: false,
       requiresThinkingAsText: false,
@@ -344,7 +251,6 @@ export const GENERATED_MODELS: ProviderModelConfig[] = [
       supportsStrictMode: false,
       supportsUsageInStreaming: true,
       thinkingFormat: "openai",
-      vercelGatewayRouting: {},
       zaiToolStream: false,
     },
     contextWindow: 131072,
@@ -371,7 +277,6 @@ export const GENERATED_MODELS: ProviderModelConfig[] = [
     name: "gpt-oss:20b",
     compat: {
       maxTokensField: "max_tokens",
-      openRouterRouting: {},
       requiresAssistantAfterToolResult: false,
       requiresReasoningContentOnAssistantMessages: false,
       requiresThinkingAsText: false,
@@ -384,7 +289,6 @@ export const GENERATED_MODELS: ProviderModelConfig[] = [
       supportsStrictMode: false,
       supportsUsageInStreaming: true,
       thinkingFormat: "openai",
-      vercelGatewayRouting: {},
       zaiToolStream: false,
     },
     contextWindow: 131072,
@@ -411,7 +315,6 @@ export const GENERATED_MODELS: ProviderModelConfig[] = [
     name: "kimi-k2.6",
     compat: {
       maxTokensField: "max_tokens",
-      openRouterRouting: {},
       requiresAssistantAfterToolResult: false,
       requiresReasoningContentOnAssistantMessages: false,
       requiresThinkingAsText: false,
@@ -424,7 +327,6 @@ export const GENERATED_MODELS: ProviderModelConfig[] = [
       supportsStrictMode: false,
       supportsUsageInStreaming: true,
       thinkingFormat: "openai",
-      vercelGatewayRouting: {},
       zaiToolStream: false,
     },
     contextWindow: 262144,
@@ -451,7 +353,6 @@ export const GENERATED_MODELS: ProviderModelConfig[] = [
     name: "kimi-k2.7-code",
     compat: {
       maxTokensField: "max_tokens",
-      openRouterRouting: {},
       requiresAssistantAfterToolResult: false,
       requiresReasoningContentOnAssistantMessages: false,
       requiresThinkingAsText: false,
@@ -464,7 +365,6 @@ export const GENERATED_MODELS: ProviderModelConfig[] = [
       supportsStrictMode: false,
       supportsUsageInStreaming: true,
       thinkingFormat: "openai",
-      vercelGatewayRouting: {},
       zaiToolStream: false,
     },
     contextWindow: 262144,
@@ -491,7 +391,6 @@ export const GENERATED_MODELS: ProviderModelConfig[] = [
     name: "kimi-k3",
     compat: {
       maxTokensField: "max_tokens",
-      openRouterRouting: {},
       requiresAssistantAfterToolResult: false,
       requiresReasoningContentOnAssistantMessages: false,
       requiresThinkingAsText: false,
@@ -504,7 +403,6 @@ export const GENERATED_MODELS: ProviderModelConfig[] = [
       supportsStrictMode: false,
       supportsUsageInStreaming: true,
       thinkingFormat: "openai",
-      vercelGatewayRouting: {},
       zaiToolStream: false,
     },
     contextWindow: 1048576,
@@ -531,7 +429,6 @@ export const GENERATED_MODELS: ProviderModelConfig[] = [
     name: "minimax-m2.7",
     compat: {
       maxTokensField: "max_tokens",
-      openRouterRouting: {},
       requiresAssistantAfterToolResult: false,
       requiresReasoningContentOnAssistantMessages: false,
       requiresThinkingAsText: false,
@@ -544,7 +441,6 @@ export const GENERATED_MODELS: ProviderModelConfig[] = [
       supportsStrictMode: false,
       supportsUsageInStreaming: true,
       thinkingFormat: "openai",
-      vercelGatewayRouting: {},
       zaiToolStream: false,
     },
     contextWindow: 196608,
@@ -571,7 +467,6 @@ export const GENERATED_MODELS: ProviderModelConfig[] = [
     name: "minimax-m3",
     compat: {
       maxTokensField: "max_tokens",
-      openRouterRouting: {},
       requiresAssistantAfterToolResult: false,
       requiresReasoningContentOnAssistantMessages: false,
       requiresThinkingAsText: false,
@@ -584,7 +479,6 @@ export const GENERATED_MODELS: ProviderModelConfig[] = [
       supportsStrictMode: false,
       supportsUsageInStreaming: true,
       thinkingFormat: "openai",
-      vercelGatewayRouting: {},
       zaiToolStream: false,
     },
     contextWindow: 512000,
@@ -611,7 +505,6 @@ export const GENERATED_MODELS: ProviderModelConfig[] = [
     name: "mistral-large-3:675b",
     compat: {
       maxTokensField: "max_tokens",
-      openRouterRouting: {},
       requiresAssistantAfterToolResult: false,
       requiresReasoningContentOnAssistantMessages: false,
       requiresThinkingAsText: false,
@@ -624,7 +517,6 @@ export const GENERATED_MODELS: ProviderModelConfig[] = [
       supportsStrictMode: false,
       supportsUsageInStreaming: true,
       thinkingFormat: "openai",
-      vercelGatewayRouting: {},
       zaiToolStream: false,
     },
     contextWindow: 262144,
@@ -643,7 +535,6 @@ export const GENERATED_MODELS: ProviderModelConfig[] = [
     name: "nemotron-3-nano:30b",
     compat: {
       maxTokensField: "max_tokens",
-      openRouterRouting: {},
       requiresAssistantAfterToolResult: false,
       requiresReasoningContentOnAssistantMessages: false,
       requiresThinkingAsText: false,
@@ -656,7 +547,6 @@ export const GENERATED_MODELS: ProviderModelConfig[] = [
       supportsStrictMode: false,
       supportsUsageInStreaming: true,
       thinkingFormat: "openai",
-      vercelGatewayRouting: {},
       zaiToolStream: false,
     },
     contextWindow: 262144,
@@ -683,7 +573,6 @@ export const GENERATED_MODELS: ProviderModelConfig[] = [
     name: "nemotron-3-super",
     compat: {
       maxTokensField: "max_tokens",
-      openRouterRouting: {},
       requiresAssistantAfterToolResult: false,
       requiresReasoningContentOnAssistantMessages: false,
       requiresThinkingAsText: false,
@@ -696,7 +585,6 @@ export const GENERATED_MODELS: ProviderModelConfig[] = [
       supportsStrictMode: false,
       supportsUsageInStreaming: true,
       thinkingFormat: "openai",
-      vercelGatewayRouting: {},
       zaiToolStream: false,
     },
     contextWindow: 262144,
@@ -723,7 +611,6 @@ export const GENERATED_MODELS: ProviderModelConfig[] = [
     name: "nemotron-3-ultra",
     compat: {
       maxTokensField: "max_tokens",
-      openRouterRouting: {},
       requiresAssistantAfterToolResult: false,
       requiresReasoningContentOnAssistantMessages: false,
       requiresThinkingAsText: false,
@@ -736,7 +623,6 @@ export const GENERATED_MODELS: ProviderModelConfig[] = [
       supportsStrictMode: false,
       supportsUsageInStreaming: true,
       thinkingFormat: "openai",
-      vercelGatewayRouting: {},
       zaiToolStream: false,
     },
     contextWindow: 262144,
@@ -758,44 +644,4 @@ export const GENERATED_MODELS: ProviderModelConfig[] = [
       xhigh: null,
     },
   },
-  {
-    id: "qwen3.5:397b",
-    name: "qwen3.5:397b",
-    compat: {
-      maxTokensField: "max_tokens",
-      openRouterRouting: {},
-      requiresAssistantAfterToolResult: false,
-      requiresReasoningContentOnAssistantMessages: false,
-      requiresThinkingAsText: false,
-      requiresToolResultName: false,
-      sendSessionAffinityHeaders: false,
-      supportsDeveloperRole: false,
-      supportsLongCacheRetention: false,
-      supportsReasoningEffort: true,
-      supportsStore: false,
-      supportsStrictMode: false,
-      supportsUsageInStreaming: true,
-      thinkingFormat: "openai",
-      vercelGatewayRouting: {},
-      zaiToolStream: false,
-    },
-    contextWindow: 262144,
-    cost: {
-      cacheRead: 0.6,
-      cacheWrite: 0,
-      input: 0.6,
-      output: 3.6,
-    },
-    input: ["text", "image"],
-    maxTokens: 65536,
-    reasoning: true,
-    thinkingLevelMap: {
-      high: null,
-      low: null,
-      medium: "medium",
-      minimal: null,
-      off: "none",
-      xhigh: null,
-    },
-  },
 ];

package/models.ts

@@ -113,18 +113,28 @@ function buildCompat(): ProviderModelConfig["compat"] {
     thinkingFormat: "openai",
     // Ollama does not support tool_choice, so strict mode is unavailable (ollama: docs.ollama.com/api/openai-compatibility, pi: types.ts#supportsStrictMode).
     supportsStrictMode: false,
-    // Anthropic cache_control not relevant; Ollama has implicit KV cache only (pi: types.ts#cacheControlFormat).
-    // Explicitly undefined: JSON.stringify drops undefined values, keeping
-    // models.generated.ts structurally consistent with assembleModels() runtime output.
     // Session affinity headers not relevant for Ollama (pi: types.ts#sendSessionAffinityHeaders).
     sendSessionAffinityHeaders: false,
     // No explicit cache-retention API (pi: types.ts#supportsLongCacheRetention).
     supportsLongCacheRetention: false,
     // Not z.ai (pi: types.ts#zaiToolStream).
     zaiToolStream: false,
+    // Anthropic cache_control not relevant; Ollama has implicit KV cache only (pi: types.ts#cacheControlFormat).
+    // Explicitly undefined: JSON.stringify drops undefined values, keeping
+    // models.generated.ts structurally consistent with assembleModels() runtime output.
     cacheControlFormat: undefined,
-    openRouterRouting: {},
-    vercelGatewayRouting: {},
+    // OpenRouter / Vercel AI Gateway routing prefs. pi-ai's OpenAI transport
+    // truthiness-checks the raw model.compat (not the resolved getCompat() value)
+    // and forwards it verbatim: `if (model.compat?.openRouterRouting)
+    // params.provider = ...` in packages/ai/src/api/openai-completions.ts. `{}` is
+    // truthy, so baking it in sent a stray `provider: {}` on every Ollama request.
+    // Express "none" as undefined, never `{}`; undefined is falsy at the wire site
+    // and JSON.stringify drops it from models.generated.ts. The boolean flags above
+    // stay explicit `false` on purpose: omitting a boolean makes getCompat() fall
+    // back to detectCompat's generic-OpenAI defaults, which are wrong for Ollama
+    // (e.g. maxTokensField -> "max_completion_tokens", supportsStrictMode -> true).
+    openRouterRouting: undefined,
+    vercelGatewayRouting: undefined,
   };
 }
 

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "pi-ollama-cloud",
-  "version": "0.12.0",
+  "version": "0.12.2",
   "type": "module",
   "keywords": [
     "pi-package"

package/pricing.generated.ts

@@ -1,7 +1,7 @@
 // Auto-generated by scripts/generate-pricing.ts
 // Do not edit manually.
-// Generated: 2026-09-12T02:41:55.206Z
-// Model count: 20
+// Generated: 2026-09-27T00:52:06.074Z
+// Model count: 17
 
 export interface ModelPrice {
   input: number;
@@ -11,11 +11,9 @@ export interface ModelPrice {
 }
 
 export const MODEL_PRICING: Record<string, ModelPrice> = {
-  "deepseek-v4-flash:0731": { input: 0.44, output: 1.32, cacheRead: 0.014, cacheWrite: 0 },
   "deepseek-v4-pro:0813": { input: 1.32, output: 3.96, cacheRead: 0.044, cacheWrite: 0 },
   "deepseek-v4.1-flash": { input: 0.3, output: 1.2, cacheRead: 0.006, cacheWrite: 0 },
   "gemma4:31b": { input: 0.14, output: 0.4, cacheRead: 0.05, cacheWrite: 0 },
-  "glm-5.1": { input: 1, output: 3.2, cacheRead: 0.2, cacheWrite: 0 },
   "glm-5.2": { input: 1.4, output: 4.4, cacheRead: 0.26, cacheWrite: 0 },
   "glm-5.3": { input: 1.4, output: 4.4, cacheRead: 0.26, cacheWrite: 0 },
   "glm-5.3-flash": { input: 0.15, output: 0.5, cacheRead: 0.03, cacheWrite: 0 },
@@ -30,5 +28,4 @@ export const MODEL_PRICING: Record<string, ModelPrice> = {
   "nemotron-3-nano:30b": { input: 0.06, output: 0.24, cacheRead: 0.06, cacheWrite: 0 },
   "nemotron-3-super": { input: 0.015, output: 0.6, cacheRead: 0.015, cacheWrite: 0 },
   "nemotron-3-ultra": { input: 0.1, output: 3, cacheRead: 0.1, cacheWrite: 0 },
-  "qwen3.5:397b": { input: 0.6, output: 3.6, cacheRead: 0.6, cacheWrite: 0 },
 };

package/reasoning.generated.ts

@@ -1,6 +1,6 @@
 // Auto-generated by scripts/generate-reasoning.ts
 // Do not edit manually.
-// Model count: 22
+// Model count: 23
 
 export type ModelsDevReasoningOption =
   | { type: "toggle" }
@@ -13,6 +13,7 @@ export const MODEL_REASONING_OPTIONS: Record<string, ModelsDevReasoningOption[]>
   "deepseek-v4-flash": [{ type: "toggle" }, { type: "effort", values: ["high", "max"] }],
   "deepseek-v4-flash:0731": [{ type: "toggle" }, { type: "effort", values: ["high", "max"] }],
   "deepseek-v4-pro": [{ type: "toggle" }, { type: "effort", values: ["high", "max"] }],
+  "deepseek-v4-pro:0813": [{ type: "toggle" }, { type: "effort", values: ["high", "max"] }],
   "deepseek-v4.1-flash": [{ type: "toggle" }, { type: "effort", values: ["low", "high", "max"] }],
   "gemma4:31b": [{ type: "toggle" }],
   "glm-5.1": [{ type: "toggle" }],