Home
Open Source Gateway GodeX New Version Links MiniMax-M3 Multimodal Power, Breaking Big Model Ecosystem Barriers

Open-source initiatives are smoothing out protocol fragmentation in the large model ecosystem. On June 1st, the open-source OpenAI Responses API gateway GodeX officially released version 1.1.0. This update not only switches the default model to the latest MiniMax-M3 but also deeply integrates multimodal understanding and reasoning control, and reconstructs Zhipu's native online search results, offering developers a more seamless local unified gateway solution.
As an open-source gateway tool, GodeX's core mission is to simplify complexity. It enables Codex, CLI tools, and developer agents to seamlessly invoke model services from DeepSeek, MiniMax, Zhipu, Xiaomi, and others—which only provide Chat Completions API—through a single "/v1/responses" protocol compatible with OpenAI. This new release means developers no longer have to worry about protocol differences across various upstream channels.
In this update, the adaptation for MiniMax-M3 is the biggest highlight. It's not just a simple name change. GodeX has supplemented multimodal input capabilities, perfectly supporting image and video understanding. Image inputs in the Responses protocol are automatically converted to MiniMax image URLs, video files are mapped to video URLs, and the system can intelligently identify and reject non-video extensions to avoid routing errors. Additionally, the new version aligns MiniMax-M3's reasoning output to the reasoning_content path, consistent with DeepSeek and Zhipu, making the reconstructed reasoning text more natural in streaming scenarios.
Regarding the balance between budget and efficiency in actual development, the new version also introduces the ability to control thinking behavior via the thinking parameter. For instance, in a video understanding scenario, enabling strong thinking mode may cause the model to consume a significant output budget on the reasoning process. By turning off thinking control through GodeX, developers can prioritize validating the video understanding path, ensuring efficient output of visible answers.
In addition to the deep restructuring of MiniMax, the Zhipu provider also received significant enhancements in this update. When Zhipu returns native web_search results, GodeX can reconstruct them into structured Responses web_search_call output. This change makes the experience more friendly for scenarios that require auditing search sources, displaying search calls, or performing secondary processing of data in agents. Currently, v1.1.0 synchronizes more complete Codex integration documentation and configuration instructions, allowing developers to easily achieve rapid deployment via npm or Docker.
Related article
Anthropic Expands Claude AI Coding Tools to Japan in Push for Overseas Growth
Anthropic, a prominent U.S. artificial intelligence firm, is intensifying its global outreach. On Wednesday, the company hosted a major developer gathering, "Code with Claude," in Tokyo, drawing close to 500 software engineers. This initiative seeks
India’s Software Giant Cuts Hiring, Promises No Layoffs as AI Agents Scale Up
As artificial intelligence reshapes traditional labor-intensive business models, Tata Consultancy Services (TCS), a premier Indian software outsourcing firm, has unveiled its strategic response. During Tuesday’s annual general meeting, the chairman o
China Locks AI Models During Gaokam to Block Instant Homework Help
With the 2026 Gaokao fast approaching, rumors regarding the suspension of AI tools during the exam period have ignited intense online debate. In response to public concern, major AI platforms and educational apps have clarified their stance: rather t
Related Special Topic Recommendations
Comments (0)
0/500

Open-source initiatives are smoothing out protocol fragmentation in the large model ecosystem. On June 1st, the open-source OpenAI Responses API gateway
As an open-source gateway tool, GodeX's core mission is to simplify complexity. It enables Codex, CLI tools, and developer agents to seamlessly invoke model services from DeepSeek, MiniMax, Zhipu, Xiaomi, and others—which only provide Chat Completions API—through a single "/v1/responses" protocol compatible with OpenAI. This new release means developers no longer have to worry about protocol differences across various upstream channels.
In this update, the adaptation for MiniMax-M3 is the biggest highlight. It's not just a simple name change. GodeX has supplemented multimodal input capabilities, perfectly supporting image and video understanding. Image inputs in the Responses protocol are automatically converted to MiniMax image URLs, video files are mapped to video URLs, and the system can intelligently identify and reject non-video extensions to avoid routing errors. Additionally, the new version aligns MiniMax-M3's reasoning output to the reasoning_content path, consistent with DeepSeek and Zhipu, making the reconstructed reasoning text more natural in streaming scenarios.
Regarding the balance between budget and efficiency in actual development, the new version also introduces the ability to control thinking behavior via the thinking parameter. For instance, in a video understanding scenario, enabling strong thinking mode may cause the model to consume a significant output budget on the reasoning process. By turning off thinking control through GodeX, developers can prioritize validating the video understanding path, ensuring efficient output of visible answers.
In addition to the deep restructuring of MiniMax, the Zhipu provider also received significant enhancements in this update. When Zhipu returns native web_search results, GodeX can reconstruct them into structured Responses web_search_call output. This change makes the experience more friendly for scenarios that require auditing search sources, displaying search calls, or performing secondary processing of data in agents. Currently, v1.1.0 synchronizes more complete Codex integration documentation and configuration instructions, allowing developers to easily achieve rapid deployment via npm or Docker.
Anthropic Expands Claude AI Coding Tools to Japan in Push for Overseas Growth
Anthropic, a prominent U.S. artificial intelligence firm, is intensifying its global outreach. On Wednesday, the company hosted a major developer gathering, "Code with Claude," in Tokyo, drawing close to 500 software engineers. This initiative seeks
India’s Software Giant Cuts Hiring, Promises No Layoffs as AI Agents Scale Up
As artificial intelligence reshapes traditional labor-intensive business models, Tata Consultancy Services (TCS), a premier Indian software outsourcing firm, has unveiled its strategic response. During Tuesday’s annual general meeting, the chairman o
China Locks AI Models During Gaokam to Block Instant Homework Help
With the 2026 Gaokao fast approaching, rumors regarding the suspension of AI tools during the exam period have ignited intense online debate. In response to public concern, major AI platforms and educational apps have clarified their stance: rather t











