@tejasa97/pi-orcarouter-provider 0.1.0 → 0.1.1

package/README.md

@@ -46,12 +46,20 @@ The static fallback includes:
 
 | Model | Reasoning | Input | Context | Max output |
 | ----- | --------- | ----- | ------- | ---------- |
-| `z-ai/glm-5.3-flash-free` | always-on | text + image | 1M | 128K |
+| `z-ai/glm-5.3-flash-free` | always-on | text + image | 110K* | 128K |
 | `z-ai/glm-5.3-flash` | always-on | text + image | 1M | 128K |
 | `z-ai/glm-5.3` | always-on | text | 1M | 128K |
 | `deepseek/deepseek-v4-flash-free` | yes | text | 164K | 32K |
 | `tencent/hy3-free` | yes | text | 128K | 32K |
 
+## Free-tier limits
+
+`z-ai/glm-5.3-flash-free` is $0 per request, but OrcaRouter gates it:
+
+- **Single-request prompt cap ~122K tokens.** The catalog advertises a 1M context, but prompts above roughly 122K tokens are rejected with `free_rate_limited` ("This prompt is longer than the free tier allows for a single request"). The provider registers the model with a 110K context window so Pi keeps sessions under the cap (Pi's system prompt and tool definitions consume part of it).
+- **Rolling usage gate.** After sustained heavy usage the same `free_rate_limited` error appears even for smaller prompts. It clears on its own after a quiet period.
+- **Transient DNS failures.** `api.orcarouter.ai` (EdgeOne CDN) intermittently fails to resolve, surfacing in Pi as "Connection error." Retrying after a minute usually works.
+
 ## GLM-5 notes
 
 Verified against the live endpoint:

package/extensions/orcarouter/models.ts

@@ -50,6 +50,15 @@ export interface RemoteModelEntry {
 
 const ZERO_COST = { input: 0, output: 0, cacheRead: 0, cacheWrite: 0 };
 
+export const FREE_GLM_FLASH_ID = "z-ai/glm-5.3-flash-free";
+
+/**
+ * The free tier rejects single requests whose prompt exceeds ~122K tokens
+ * (measured empirically; the catalog claims 1M). pi adds its own system
+ * prompt and tool definitions on top, so the honest usable window is lower.
+ */
+export const FREE_TIER_CONTEXT_WINDOW = 110_000;
+
 /** Model ids that are clearly not chat-completions models. */
 const NON_CHAT_ID = /(image|video|kling|dub|tts|speech|sora|whisper|embed|dall|flux|veo|seedance|stable|diffusion|audio|voice|music|rerank)/i;
 
@@ -254,9 +263,11 @@ function buildModel(id: string, entry: RemoteModelEntry | undefined, baseUrl: st
 		cost: buildCost(entry?.pricing),
 		compat: buildCompat(reasoning),
 		contextWindow:
-			numberValue(entry?.context_length) ??
-			numberValue(entry?.top_provider?.context_length) ??
-			defaultContextWindow(id),
+			id === FREE_GLM_FLASH_ID
+				? FREE_TIER_CONTEXT_WINDOW
+				: numberValue(entry?.context_length) ??
+					numberValue(entry?.top_provider?.context_length) ??
+					defaultContextWindow(id),
 		maxTokens:
 			numberValue(entry?.max_completion_tokens) ??
 			numberValue(entry?.top_provider?.max_completion_tokens) ??
@@ -269,7 +280,7 @@ function buildModel(id: string, entry: RemoteModelEntry | undefined, baseUrl: st
  * /v1/models fetch fails. The free GLM 5.3 Flash tier is the headline model.
  */
 export const STATIC_MODEL_IDS = [
-	"z-ai/glm-5.3-flash-free",
+	FREE_GLM_FLASH_ID,
 	"z-ai/glm-5.3-flash",
 	"z-ai/glm-5.3",
 	"deepseek/deepseek-v4-flash-free",
@@ -280,13 +291,11 @@ export function buildStaticModels(baseUrl: string): readonly OrcaRouterModel[] {
 	return STATIC_MODEL_IDS.map((id) =>
 		buildModel(
 			id,
-			id === "z-ai/glm-5.3-flash-free"
+			id === FREE_GLM_FLASH_ID
 				? {
 						id,
 						supported_endpoint_types: null,
 						name: "Z.ai: GLM 5.3 Flash (Free)",
-						context_length: 1_000_000,
-						max_completion_tokens: 128_000,
 						architecture: { input_modalities: ["text", "image", "video"] },
 						pricing: { request: "0.000000" },
 					}

package/package.json

@@ -1,6 +1,6 @@
 {
 	"name": "@tejasa97/pi-orcarouter-provider",
-	"version": "0.1.0",
+	"version": "0.1.1",
 	"description": "Pi provider package for OrcaRouter (GLM 5.3 Flash Free, DeepSeek, Kimi, and more) — OpenAI-compatible gateway with dynamic model discovery.",
 	"type": "module",
 	"engines": {