pi-nvidia-nim 1.1.15 → 1.1.17

package/README.md

@@ -1,6 +1,6 @@
 # pi-nvidia-nim
 
-NVIDIA NIM API provider extension for [pi coding agent](https://github.com/badlogic/pi-mono) - access 100+ models from [build.nvidia.com](https://build.nvidia.com) including DeepSeek V3.2, Kimi K2.5, MiniMax M2.1, GLM-4.7, Qwen3, Llama 4, and many more.
+NVIDIA NIM API provider extension for [pi coding agent](https://github.com/badlogic/pi-mono) - access 100+ models from [build.nvidia.com](https://build.nvidia.com) including DeepSeek V3.2, Kimi K2.5, MiniMax M2.1, GLM-5, GLM-4.7, Qwen3, Llama 4, and many more.
 
 https://github.com/user-attachments/assets/f44773e4-9bf8-4bb5-a9c0-d5938030701c
 
@@ -73,10 +73,10 @@ When you change the thinking level in pi (`Shift+Tab` to cycle), the extension:
 1. **Maps `"minimal"` → `"low"`** - NIM only accepts `low`, `medium`, `high` (not `minimal`). Selecting "minimal" in pi works fine; it's silently mapped.
 2. **Injects `chat_template_kwargs`** per model to actually enable thinking:
    - DeepSeek V3.x, R1 distills: `{ thinking: true }`
-   - GLM-4.7: `{ enable_thinking: true, clear_thinking: false }`
+   - GLM-5, GLM-4.7: `{ enable_thinking: true, clear_thinking: false }`
    - Kimi K2.5, K2-thinking: `{ thinking: true }`
    - Qwen3, QwQ: `{ enable_thinking: true }`
-3. **Explicitly disables thinking** when the level is "off" for models that think by default (e.g., GLM-4.7).
+3. **Explicitly disables thinking** when the level is "off" for models that think by default (e.g., GLM-5, GLM-4.7).
 4. **Uses `system` role** instead of `developer` for all NIM models - the `developer` role combined with `chat_template_kwargs` causes 500 errors on NIM.
 
 ### Supported thinking levels
@@ -91,7 +91,7 @@ When you change the thinking level in pi (`Shift+Tab` to cycle), the extension:
 
 ## Available Models
 
-The extension ships with curated metadata for 38 featured models. At startup, it also queries the NVIDIA NIM API to discover additional models automatically.
+The extension ships with curated metadata for 39 featured models. At startup, it also queries the NVIDIA NIM API to discover additional models automatically.
 
 ### Featured Models
 
@@ -102,6 +102,7 @@ The extension ships with curated metadata for 38 featured models. At startup, it
 | `moonshotai/kimi-k2.5` | ✅ | | 256K |
 | `moonshotai/kimi-k2-thinking` | ✅ | | 128K |
 | `minimaxai/minimax-m2.1` | | | 1M |
+| `z-ai/glm5` | ✅ | | 128K |
 | `z-ai/glm4.7` | ✅ | | 128K |
 | `openai/gpt-oss-120b` | | | 128K |
 | `qwen/qwen3-coder-480b-a35b-instruct` | ✅ | | 256K |
@@ -120,7 +121,7 @@ The extension ships with curated metadata for 38 featured models. At startup, it
 
 ### Tool Calling
 
-All major models support OpenAI-compatible tool calling. Tested and confirmed working with DeepSeek V3.2, GLM-4.7, Qwen3, Kimi K2.5, and others.
+All major models support OpenAI-compatible tool calling. Tested and confirmed working with DeepSeek V3.2, GLM-5, GLM-4.7, Qwen3, Kimi K2.5, and others.
 
 ## How It Works
 

package/index.ts

@@ -22,14 +22,14 @@
  * parameters:
  *
  * - DeepSeek V3.x: `chat_template_kwargs: { thinking: true }`
- * - GLM-4.7:       `chat_template_kwargs: { enable_thinking: true, clear_thinking: false }`
+ * - GLM-5/4.7:     `chat_template_kwargs: { enable_thinking: true, clear_thinking: false }`
  * - Kimi K2.5:     `chat_template_kwargs: { thinking: true }` (also accepts reasoning_effort)
  * - Qwen3:         `chat_template_kwargs: { enable_thinking: true }`
  *
  * NIM only accepts `reasoning_effort` values of "low", "medium", "high" - NOT
  * "minimal". The extension maps pi's "minimal" level to "low" automatically.
  *
- * Some models (e.g., GLM-4.7) always produce reasoning output regardless of
+ * Some models (e.g., GLM-5, GLM-4.7) always produce reasoning output regardless of
  * thinking settings.
  */
 
@@ -60,7 +60,7 @@ const PROVIDER_NAME = "nvidia-nim";
  * When a user enables thinking in pi (any level > off), we inject these kwargs
  * into the request body. Models not listed here either:
  * - Don't support thinking (non-reasoning models)
- * - Always think regardless (GLM-4.7 without explicit kwargs)
+ * - Always think regardless (GLM models without explicit kwargs)
  * - Work with standard reasoning_effort (rare on NIM)
  */
 interface ThinkingConfig {
@@ -102,11 +102,15 @@ const THINKING_CONFIGS: Record<string, ThinkingConfig> = {
 		enableKwargs: { thinking: true },
 		disableKwargs: { thinking: false },
 	},
-	// GLM-4.7 always thinks by default, but can be controlled
+	// GLM models (Z-AI) - think by default, but can be controlled
 	"z-ai/glm4.7": {
 		enableKwargs: { enable_thinking: true, clear_thinking: false },
 		disableKwargs: { enable_thinking: false },
 	},
+	"z-ai/glm5": {
+		enableKwargs: { enable_thinking: true, clear_thinking: false },
+		disableKwargs: { enable_thinking: false },
+	},
 	// Kimi models: chat_template_kwargs works, reasoning_effort also works
 	"moonshotai/kimi-k2.5": {
 		enableKwargs: { thinking: true },
@@ -245,6 +249,7 @@ const CONTEXT_WINDOWS: Record<string, number> = {
 	// MiniMax
 	"minimaxai/minimax-m2": 1048576,
 	"minimaxai/minimax-m2.1": 1048576,
+	"minimaxai/minimax-m2.7": 204800,
 	// Meta Llama
 	"meta/llama-3.1-405b-instruct": 131072,
 	"meta/llama-3.1-70b-instruct": 131072,
@@ -320,6 +325,7 @@ const CONTEXT_WINDOWS: Record<string, number> = {
 	"openai/gpt-oss-20b": 131072,
 	// Z-AI / GLM
 	"z-ai/glm4.7": 131072,
+	"z-ai/glm5": 131072,
 	// StepFun
 	"stepfun-ai/step-3.5-flash": 131072,
 	// ByteDance
@@ -368,9 +374,11 @@ const MAX_TOKENS: Record<string, number> = {
 	"moonshotai/kimi-k2-thinking": 16384,
 	"minimaxai/minimax-m2": 8192,
 	"minimaxai/minimax-m2.1": 8192,
+	"minimaxai/minimax-m2.7": 8192,
 	"meta/llama-4-maverick-17b-128e-instruct": 16384,
 	"meta/llama-4-scout-17b-16e-instruct": 16384,
 	"z-ai/glm4.7": 16384,
+	"z-ai/glm5": 16384,
 	"qwen/qwen3-coder-480b-a35b-instruct": 65536,
 	"nvidia/llama-3.1-nemotron-ultra-253b-v1": 32768,
 	"openai/gpt-oss-120b": 16384,
@@ -394,6 +402,8 @@ const FEATURED_MODELS = [
 	"moonshotai/kimi-k2-instruct-0905",
 	"minimaxai/minimax-m2.1",
 	"minimaxai/minimax-m2",
+	"minimaxai/minimax-m2.7",
+	"z-ai/glm5",
 	"z-ai/glm4.7",
 	"openai/gpt-oss-120b",
 	"openai/gpt-oss-20b",
@@ -494,7 +504,7 @@ function nimStreamSimple(
 					// Inject chat_template_kwargs to enable thinking
 					p.chat_template_kwargs = thinkingConfig.enableKwargs;
 				} else if (thinkingConfig.disableKwargs) {
-					// Explicitly disable thinking (some models think by default, e.g. GLM-4.7)
+					// Explicitly disable thinking (some models think by default, e.g. GLM-5/4.7)
 					p.chat_template_kwargs = thinkingConfig.disableKwargs;
 				}
 			}

package/package.json

@@ -1,6 +1,6 @@
 {
   "name": "pi-nvidia-nim",
-  "version": "1.1.15",
+  "version": "1.1.17",
   "description": "NVIDIA NIM API provider extension for pi coding agent — access 100+ models from build.nvidia.com",
   "type": "module",
   "files": [