diff --git a/__init__.py b/__init__.py
index f5fa3bb..cc29b36 100644
--- a/__init__.py
+++ b/__init__.py
@@ -3,7 +3,7 @@
     "name": "ComfyUI-ThinkingLLM",
     "description": "A multimodal ComfyUI AI node with Qwen3.5, Qwen3-VL, Qwen2.5-VL, Qwen3, and Gemma 4 integrations. Features live thinking in the terminal to see what the LLM is doing in real time.",
     "author": "goodguy1963",
-    "version": "2.3.24",
+    "version": "2.4.0",
     "url": "https://github.com/goodguy1963/ComfyUI-ThinkingLLM",
     "category": "image"
 }
diff --git a/custom_nodes.json b/custom_nodes.json
index 5e4785f..6f50008 100644
--- a/custom_nodes.json
+++ b/custom_nodes.json
@@ -9,5 +9,5 @@
   "description": "A multimodal ComfyUI AI node with Qwen3.5, Qwen3-VL, Qwen2.5-VL, Qwen3, and Gemma 4 integrations. Features live thinking in the terminal to see what the LLM is doing in real time.",
   "category": "image",
   "tags": ["vision", "language", "multimodal", "thinkingllm", "qwen", "gguf", "flash_attention", "performance", "wan2.2", "video", "i2v", "t2v", "storyboard", "caching", "fork", "gpl-3.0"],
-  "version": "2.3.24"
+  "version": "2.4.0"
 }
diff --git a/pyproject.toml b/pyproject.toml
index c07a1d0..9cbd629 100644
--- a/pyproject.toml
+++ b/pyproject.toml
@@ -1,7 +1,7 @@
 [project]
 name = "ComfyUI-ThinkingLLM"
 description = "A multimodal ComfyUI AI node with Qwen3.5, Qwen3-VL, Qwen2.5-VL, Qwen3, and Gemma 4 integrations. Features live thinking in the terminal to see what the LLM is doing in real time."
-version = "2.3.24"
+version = "2.4.0"
 license = {file = "LICENSE"}
 dependencies = ["transformers>=5.2.0", "torch", "huggingface-hub", "hf_xet", "psutil", "numpy", "Pillow", "opencv-python", "bitsandbytes", "accelerate", "sentencepiece", "tiktoken", "faster-whisper", "llama-cpp-python"]
 
diff --git a/update.md b/update.md
index c240fef..f2b442f 100644
--- a/update.md
+++ b/update.md
@@ -1,5 +1,20 @@
 # ComfyUI-QwenVL Update Log
 
+## Version 2.4.0 (2026/08/12) - Duration-Aware Video Prompting and QwenVL Core
+
+### 2.4.0 Changes
+
+- Added duration-aware LTX 2.3 and MiniMax H3 prompting across the Hugging Face, ComfyUI-native, and GGUF video workflows.
+- Split MiniMax H3 prompting into dedicated T2VA, I2VA, FL2VA, L2VA, and full-reference contracts with frame-grid normalization and safer best-effort output cleanup.
+- Refactored the QwenVL implementation into focused shared modules for model access, runtime configuration, prompt state, prompt contracts, and media handling while preserving compatibility exports.
+- Improved streaming so reasoning stays separate from final content, unfinished thinking traces trigger a direct finalization retry, and terminal-only punctuation remains safe on legacy Windows code pages.
+- Expanded preset recommendations, tooltips, system prompts, and documentation for video token budgets, thinking behavior, reference roles, dialogue fidelity, and duration handling.
+
+### 2.4.0 Validation
+
+- Expanded regression coverage for compatibility, downloads, caching, duration-aware prompting, MiniMax structures, streaming, and node registration.
+- All 131 isolated regression tests pass in the bundled ComfyUI Python runtime.
+
 ## Version 2.3.24 (2026/08/08) - System Prompt Preset Utility
 
 ### 2.3.24 Changes