@ssweens/pi-vertex 1.1.9 → 1.1.12

package/CHANGELOG.md

@@ -2,6 +2,15 @@
 
 All notable changes to this project will be documented in this file.
 
+## [1.1.12] - 2026-06-03
+### Added
+- **Claude Opus 4.8** (`claude-opus-4-8`) — added the new Vertex/Gemini Enterprise Agent Platform Anthropic partner model with 1M context, 128K max output tokens, and the same published global/regional pricing tier used by Opus 4.7.
+- **Grok 4.3** (`grok-4.3`) — added the new Vertex/Gemini Enterprise Agent Platform xAI partner model with 200K context, 32,768 max output tokens (conservative default), and the published Google pricing tier for Grok 4.3.
+
+## [1.1.10] - 2026-05-28
+### Fixed
+- Switched `registerProvider("vertex")` `apiKey` from legacy env-var shorthand (`"GOOGLE_CLOUD_PROJECT"`) to explicit env syntax (`"$GOOGLE_CLOUD_PROJECT"`). This removes deprecation warnings and future-proofs provider registration against upcoming removal of implicit env-var detection.
+
 ## [1.1.9] - 2026-05-19
 ### Fixed
 - **Gemini 3/3.5 default thinking preserved** — previously, if no reasoning effort level was specified, we silently overrode Gemini 3/3.5 models to use their lowest thinking level (`MINIMAL`). This severely degraded the intelligence of Gemini 3.5 Flash (defaulting it to no real reasoning/thinking) and Gemini 3 Pro/Flash. We now completely omit `thinkingConfig` when `options.reasoning` is undefined, allowing Gemini 3/3.5 models to use their native GA defaults (`MEDIUM` for 3.5 Flash, `HIGH` for others).

package/README.md

@@ -15,11 +15,11 @@ Set your GCP project and credentials. Vertex AI models (Gemini, Claude, Llama, D
 
 ## Features
 
-- **48 models** across 4 categories:
+- **45 models** across 4 categories:
   - **Gemini** (9): 3.1 Pro, 3.1 Flash-Lite, 3 Flash, 2.5 Pro, 2.5 Flash, 2.5 Flash-Lite, 2.0 Flash, 2.0 Flash-Lite
-  - **Claude** (10): Opus 4.7, Opus 4.6, Sonnet 4.6, Opus/Sonnet/Haiku 4.5, Opus 4.1, Opus 4, Sonnet 4, 3.5 Sonnet v2
+  - **Claude** (11): Opus 4.8, Opus 4.7, Opus 4.6, Sonnet 4.6, Opus/Sonnet/Haiku 4.5, Opus 4.1, Opus 4, Sonnet 4, 3.5 Sonnet v2
   - **Llama** (3): 4 Maverick, 4 Scout, 3.3 70B
-  - **Other MaaS** (26): Grok, Gemma, Mistral, DeepSeek, Qwen, OpenAI GPT-OSS, Kimi, MiniMax, GLM
+  - **Other MaaS** (22): Grok 4.3, Grok 4.20, Grok 4.1 Fast, Gemma, Mistral, DeepSeek, Qwen, OpenAI GPT-OSS, Kimi, MiniMax, GLM
 
 - **Unified streaming**: Single provider, multiple model families
 - **Full tool calling support**: All models with multi-turn tool use and proper tool result handling
@@ -86,7 +86,8 @@ pi --provider vertex --model gemini-2.5-pro --version
 
 ```bash
 # Use any supported model
-pi --provider vertex --model claude-opus-4-6
+pi --provider vertex --model claude-opus-4-8
+pi --provider vertex --model grok-4.3
 pi --provider vertex --model gemini-2.5-pro
 pi --provider vertex --model llama-4-maverick
 pi --provider vertex --model deepseek-v3.2
@@ -100,8 +101,11 @@ pi --provider vertex --model claude-sonnet-4-6 --reasoning high
 Add to `~/.bashrc` or `~/.zshrc`:
 
 ```bash
-# Claude 4.6 Opus
-alias pic="GOOGLE_CLOUD_PROJECT=your-project pi --provider vertex --model claude-opus-4-6"
+# Claude 4.8 Opus
+alias pic="GOOGLE_CLOUD_PROJECT=your-project pi --provider vertex --model claude-opus-4-8"
+
+# Grok 4.3
+alias pig4="GOOGLE_CLOUD_PROJECT=your-project pi --provider vertex --model grok-4.3"
 
 # Gemini 2.5 Pro
 alias pig="GOOGLE_CLOUD_PROJECT=your-project pi --provider vertex --model gemini-2.5-pro"
@@ -132,6 +136,7 @@ Prices shown are for the **global** endpoint. Non-global regions (us-east5, euro
 
 | Model | Context | Max Tokens | Input | Reasoning | Price global (in/out) | Price regional (in/out) |
 |-------|---------|------------|-------|-----------|----------------------|------------------------|
+| claude-opus-4-8 | 1M | 128,000 | text, image | ✅ | $5.00/$25.00 | $5.50/$27.50 |
 | claude-opus-4-7 | 1M | 128,000 | text, image | ✅ | $5.00/$25.00 | $5.50/$27.50 |
 | claude-opus-4-6 | 1M | 128,000 | text, image | ✅ | $5.00/$25.00 | $5.50/$27.50 |
 | claude-sonnet-4-6 | 1M | 128,000 | text, image | ✅ | $3.00/$15.00 | $3.30/$16.50 |
@@ -173,8 +178,9 @@ Prices shown are for the **global** endpoint. Non-global regions (us-east5, euro
 | minimax-m2 | 196K | minimaxai | $0.30/$1.20 | global |
 | glm-5 | 200K | zai-org | $1.00/$3.20 | global |
 | glm-4.7 | 200K | zai-org | $0.60/$2.20 | global |
-| grok-4.20-reasoning | 200K | xai | $1.25/$2.50 | global |
-| grok-4.1-fast-reasoning | 128K | xai | $0.20/$0.50 | global |
+| grok-4.3 | 1M | xai | $1.25/$2.50 | global |
+| grok-4.20-reasoning | 1M | xai | $1.25/$2.50 | global |
+| grok-4.1-fast-reasoning | 2M | xai | $0.20/$0.50 | global |
 | gemma-4-26b-a4b-it | 262K | google | $0.15/$0.60 | global |
 
 ## Regional Endpoints

package/TEST_COVERAGE.md

@@ -7,6 +7,7 @@
 ## Manual Verification
 - Claude 4.6 streaming verified via Anthropic Vertex SDK.
 - Mid-session model switching (tool call replay) verified interactively in pi.
+- Registry sanity checks verified locally for `claude-opus-4-8` and `grok-4.3` after model additions.
 
 ## Gaps / Next Steps
 - Add automated integration tests for Anthropic Vertex streaming and tool replay.

package/index.ts

@@ -21,7 +21,8 @@
  *     GOOGLE_APPLICATION_CREDENTIALS          (optional, for service account auth)
  *
  * Usage:
- *   pi --provider vertex --model claude-opus-4-6
+ *   pi --provider vertex --model claude-opus-4-8
+ *   pi --provider vertex --model grok-4.3
  *   pi --provider vertex --model gemini-2.5-pro
  *   pi --provider vertex --model llama-4-maverick
  */
@@ -90,8 +91,8 @@ export default function (pi: ExtensionAPI) {
     // Use a placeholder baseUrl (actual URLs built per-request based on model region)
     baseUrl: "https://aiplatform.googleapis.com",
 
-    // Use env var name for detection
-    apiKey: "GOOGLE_CLOUD_PROJECT",
+    // Use explicit env var reference
+    apiKey: "$GOOGLE_CLOUD_PROJECT",
 
     // API type varies by model
     api: "vertex-unified",

package/models/claude.ts

@@ -11,6 +11,33 @@
 import type { VertexModelConfig } from "../types.js";
 
 export const CLAUDE_MODELS: VertexModelConfig[] = [
+  // Claude 4.8 series
+  {
+    id: "claude-opus-4-8",
+    name: "Claude Opus 4.8",
+    apiId: "claude-opus-4-8",
+    publisher: "anthropic",
+    endpointType: "maas",
+    contextWindow: 1000000,
+    maxTokens: 128000,
+    input: ["text", "image"],
+    reasoning: true,
+    tools: true,
+    cost: {
+      input: 5.00,
+      output: 25.00,
+      cacheRead: 0.50,
+      cacheWrite: 6.25,
+    },
+    costRegional: {
+      input: 5.50,
+      output: 27.50,
+      cacheRead: 0.55,
+      cacheWrite: 6.875,
+    },
+    region: "global",
+  },
+
   // Claude 4.7 series
   {
     id: "claude-opus-4-7",

package/models/maas.ts

@@ -9,13 +9,32 @@ import type { VertexModelConfig } from "../types.js";
 
 export const MAAS_MODELS: VertexModelConfig[] = [
   // --- xAI Grok ---
+  {
+    id: "grok-4.3",
+    name: "Grok 4.3",
+    apiId: "grok-4.3",
+    publisher: "xai",
+    endpointType: "maas",
+    contextWindow: 1000000,
+    maxTokens: 32768,
+    input: ["text", "image"],
+    reasoning: true,
+    tools: true,
+    cost: {
+      input: 1.25,
+      output: 2.50,
+      cacheRead: 0.20,
+      cacheWrite: 0,
+    },
+    region: "global",
+  },
   {
     id: "grok-4.20-reasoning",
     name: "Grok 4.20 Reasoning",
     apiId: "grok-4.20-reasoning",
     publisher: "xai",
     endpointType: "maas",
-    contextWindow: 200000,
+    contextWindow: 1000000,
     maxTokens: 32000,
     input: ["text", "image"],
     reasoning: true,
@@ -34,7 +53,7 @@ export const MAAS_MODELS: VertexModelConfig[] = [
     apiId: "grok-4.1-fast-reasoning",
     publisher: "xai",
     endpointType: "maas",
-    contextWindow: 128000,
+    contextWindow: 2000000,
     maxTokens: 32000,
     input: ["text", "image"],
     reasoning: true,

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "@ssweens/pi-vertex",
-  "version": "1.1.9",
+  "version": "1.1.12",
   "description": "Google Vertex AI provider for Pi coding agent - supports Gemini, Claude, and all MaaS models",
   "type": "module",
   "main": "index.ts",