> ## Documentation Index > Fetch the complete documentation index at: https://docs.aihubmix.com/llms.txt > Use this file to discover all available pages before exploring further. # Modèles d'IA gratuits sur AIHubMix > Modèles d'IA gratuits : le guide ultime pour construire avec une IA à coût zéro en 2026 sur AIHubMix Modèles d'IA gratuits sur AIHubMix : accès gratuit aux grands modèles populaires **Les API d'IA gratuites sont le moyen le plus rapide d'expédier des fonctionnalités d'IA en 2026** — mais la plupart des plateformes « gratuites » exigent une carte de crédit, font expirer les essais ou imposent des plafonds d'utilisation surprises. AIHubMix adopte une approche différente : une passerelle unifiée compatible OpenAI avec **plus de 27 modèles LLM et de génération d'images réellement gratuits** subventionnés par la plateforme, notamment GPT-5.5 et GPT-Image-2 d'OpenAI, Gemini 3 de Google, Zhipu GLM-5.1, Kimi, MiniMax et Xiaomi MiMo. Pas de carte de crédit. Pas d'expiration d'essai. Une seule clé API, tous les grands modèles. ## 🚀 Dernière mise à jour : GPT-5.5 et GPT-Image-2 sont désormais gratuits AIHubMix s'engage à offrir une valeur maximale à ses utilisateurs. Dans cette mise à jour, les versions gratuites de deux des derniers modèles phares d'OpenAI — **GPT-5.5 et GPT-Image-2** — sont officiellement disponibles. L'API officielle d'OpenAI n'offrant pas d'accès gratuit à ces modèles, AIHubMix **continue d'investir pour subventionner les coûts d'inférence**, abaissant à zéro la barrière d'entrée aux modèles de premier plan. [**GPT-5.5-free**](https://aihubmix.com/model/gpt-5.5-free) Une mise à niveau complète en matière de profondeur de raisonnement, d'orchestration d'agent, d'utilisation d'outils, de génération de code et d'analyse de données — actuellement le modèle disponible le plus performant d'OpenAI. L'accès gratuit sur AIHubMix est le moyen le plus rapide de comparer GPT-5.5 à Claude Opus 4.6, Gemini 3.1 Pro et GLM-5.1 sans payer au jeton. Exemples d'utilisation de l'API [**GPT-5.5-free**](https://aihubmix.com/model/gpt-5.5-free) ```python Python theme={null} import openai client = openai.OpenAI( api_key="", # Replace with the key generated in AIHubMix base_url="https://aihubmix.com/v1" ) response = client.chat.completions.create( model="gpt-5.5-free", # The reasoning depth of the model defaults to medium messages=[ {"role": "user", "content": "Hello, how are you?"} ], temperature=0.7 # Default is 1 ) print(response.choices[0].message.content) ``` ```python Responses-API theme={null} from openai import OpenAI client = OpenAI( api_key="", base_url="https://aihubmix.com/v1" ) response = client.responses.create( model="gpt-5.5-free", input="Hello, how are you?" ) print(response.output_text) ``` ```typescript TypeScript theme={null} import OpenAI from "openai"; const client = new OpenAI({ apiKey: "", baseURL: "https://aihubmix.com/v1", }); const response = await client.chat.completions.create({ model: "gpt-5.5-free", messages: [{ role: "user", content: "Hello, how are you?" }], temperature: 0.7, }); console.log(response.choices[0].message.content); ``` ```bash Curl theme={null} curl https://aihubmix.com/v1/chat/completions \ -H "Content-Type: application/json" \ -H "Authorization: Bearer " \ -d '{ "model": "gpt-5.5-free", "messages": [{"role": "user", "content": "Hello, how are you?"}], "temperature": 0.7 }' ``` [**GPT-Image-2-free**](https://aihubmix.com/model/gpt-image-2-free) Product photography, posters, avatars, illustrations, e-commerce assets, social media graphics, livestream thumbnails — all mainstream image generation use cases covered in one call, with output quality at commercial-grade level. The first OpenAI image model with built-in reasoning and \~99% character-level text rendering accuracy across Latin, CJK, Hindi, and other scripts. **API Usage Examples** ```python Python_generate theme={null} import base64 from openai import OpenAI client = OpenAI( api_key="", # Replace with the key generated in AIHubMix base_url="https://aihubmix.com/v1" ) response = client.images.generate( model="gpt-image-2-free", prompt="A vase of flowers on a table, with intense contrasting colors and thick, expressive brushstrokes. Render the image so it looks painted in Fauvist style.", n=1, # Number of images to generate, supports 1-10 size="auto", # Image size: 1024x1024, 1024x1536, 1536x1024, 4096x4096, auto (default) quality="auto" # Image quality: high, medium, low, auto (default) ) image_bytes = base64.b64decode(response.data[0].b64_json) with open("output.png", "wb") as f: f.write(image_bytes) ``` ```python Python_edit theme={null} import base64 from openai import OpenAI client = OpenAI( api_key="", base_url="https://aihubmix.com/v1" ) # Pass one or more reference images plus an editing prompt response = client.images.edit( model="gpt-image-2-free", image=[open("product.png", "rb")], # Reference image(s) to edit / extend prompt="Change the background to a sunset beach scene, keep the product centered and unchanged.", n=1, size="1024x1024", quality="high" ) image_bytes = base64.b64decode(response.data[0].b64_json) with open("edited.png", "wb") as f: f.write(image_bytes) ``` ```bash Curl theme={null} curl https://aihubmix.com/v1/images/generations \ -H "Content-Type: application/json" \ -H "Authorization: Bearer " \ -d '{ "model": "gpt-image-2-free", "prompt": "A vase of flowers on a table, with intense contrasting colors and thick, expressive brushstrokes. Render the image so it looks painted in Fauvist style.", "n": 1, "size": "auto", "quality": "auto" }' ``` ```python Python_response theme={null} import base64 from openai import OpenAI client = OpenAI( api_key="", base_url="https://aihubmix.com/v1" ) # Generate images via the Responses API with the image_generation tool response = client.responses.create( model="gpt-5.5-free", input="Generate an image of a vase of flowers on a table, Fauvist style with intense contrasting colors and thick, expressive brushstrokes.", tools=[{"type": "image_generation"}], ) for item in response.output: if item.type == "image_generation_call": image_bytes = base64.b64decode(item.result) with open("output.png", "wb") as f: f.write(image_bytes) break ``` *** ## Why Use Free AI APIs in 2026? Free AI model APIs unlock four concrete benefits that paid-only access cannot match: * **Side-by-side model evaluation** — Compare GPT-5.5, Claude Opus 4.6, Gemini 3.1 Pro, GLM-5.1, and Kimi on the same prompts before committing to a paid plan. * **Zero-cost prototyping** — Build proof-of-concept agents, chatbots, and automation pipelines without burning your credit card during the discovery phase. * **Cost-aware production routing** — Route low-stakes traffic (batch summarization, log analysis, draft generation) to free models while reserving paid quota for revenue-critical paths. * **Hobbyist and student access** — Indie developers, students, and side-project builders gain access to frontier models that would otherwise cost hundreds per month. The catch with most "free LLM API" providers is fragmentation: Google AI Studio gives you Gemini, Groq gives you Llama, OpenRouter gives you a different mix every week, and each requires a separate account, API key, and rate-limit strategy. AIHubMix consolidates 27+ free models behind **one OpenAI-compatible endpoint** with automatic provider failover — drop-in replacement for any existing OpenAI SDK call. *** ## Complete Free Model Catalog (27+ Models, May 2026) AIHubMix currently offers 27+ free models spanning major providers including OpenAI, Google, Zhipu, Kimi, MiniMax, and Xiaomi — and the lineup keeps growing as new models ship. ### General-Purpose Chat & Reasoning Models Covering the GPT-4o and GPT-4.1 families plus Gemini Flash and domestic flagships — suited for everyday Q\&A, content generation, document analysis, and multilingual chat. `gpt-4o-free` supports mixed text-and-image input, `gemini-3-flash-preview-free` offers ultra-long context (1M+ tokens), and the rest balance speed and capability differently. | Model | Context | Highlights | | :------------------------------------------------------------------------------------ | :------ | :-------------------------------------------------- | | [gpt-4o-free](https://aihubmix.com/model/gpt-4o-free) | 128K | Multimodal, vision-capable | | [gpt-4.1-free](https://aihubmix.com/model/gpt-4.1-free) | 1M | Complex instruction following, long-form generation | | [gpt-4.1-mini-free](https://aihubmix.com/model/gpt-4.1-mini-free) | 1M | Balanced speed and capability | | [gpt-4.1-nano-free](https://aihubmix.com/model/gpt-4.1-nano-free) | 1M | Lightweight, high-frequency tasks | | [gemini-3-flash-preview-free](https://aihubmix.com/model/gemini-3-flash-preview-free) | 1M+ | Ultra-long context, multimodal input | | [glm-4.7-flash-free](https://aihubmix.com/model/glm-4.7-flash-free) | 128K | Fast response, multilingual support | | [mimo-v2-flash-free](https://aihubmix.com/model/mimo-v2-flash-free) | 128K | Low-latency conversation | | [ling-2.6-flash-free](https://aihubmix.com/model/ling-2.6-flash-free) | 128K | Strong contextual coherence | ### Free Coding Models (Largest Category) The deepest category in the free tier — bringing together specialized coding model series from Kimi, MiniMax, Zhipu GLM, and Qwen. If you're searching for a **free GitHub Copilot alternative** or a **free Cursor backend**, this is where to start. | Model | Strength | | :------------------------------------------------------------------------------ | :----------------------------------------- | | [kimi-for-coding-free](https://aihubmix.com/model/kimi-for-coding-free) | Multi-file context, refactoring, debugging | | [k2.6-code-preview-free](https://aihubmix.com/model/k2.6-code-preview-free) | Algorithmic and systems-level code | | [coding-minimax-m2-free](https://aihubmix.com/model/coding-minimax-m2-free) | MiniMax coding series | | [coding-minimax-m2.1-free](https://aihubmix.com/model/coding-minimax-m2.1-free) | MiniMax coding series | | [coding-minimax-m2.5-free](https://aihubmix.com/model/coding-minimax-m2.5-free) | MiniMax coding series | | [coding-minimax-m2.7-free](https://aihubmix.com/model/coding-minimax-m2.7-free) | Latest MiniMax coding release | | [coding-glm-4.6-free](https://aihubmix.com/model/coding-glm-4.6-free) | GLM coding series | | [coding-glm-4.7-free](https://aihubmix.com/model/coding-glm-4.7-free) | GLM coding series | | [coding-glm-5-free](https://aihubmix.com/model/coding-glm-5-free) | GLM-5, 745B MoE, Claude Opus 4.5 parity | | [coding-glm-5-turbo-free](https://aihubmix.com/model/coding-glm-5-turbo-free) | GLM coding accelerated edition | | [coding-glm-5.1-free](https://aihubmix.com/model/coding-glm-5.1-free) | **#1 on SWE-bench Pro (58.4%)** | | [step-3.5-flash-free](https://aihubmix.com/model/step-3.5-flash-free) | Lightweight completion, low latency | ### Free Image Generation Models [**GPT-Image-2-free**](https://aihubmix.com/model/gpt-image-2-free) OpenAI's next-generation image generation model released in April 2026, and its first image model with built-in reasoning. Before generating, it automatically plans composition, retrieves visual references from the web, and self-checks output — yielding noticeably better quality than GPT Image 1.5. Supports up to **4096×4096 resolution**, generates roughly 2× faster than GPT Image 1.5, and produces up to 8 stylistically consistent images from a single prompt. Text rendering is a particular strength — covering Latin, CJK, Hindi, and other scripts with character-level accuracy of about 99%, making it ideal for posters, marketing assets, UI prototypes, and any scenario requiring precise typography. [**gemini-3.1-flash-image-preview-free**](https://aihubmix.com/model/gemini-3.1-flash-image-preview-free) (Nano Banana 2) Released by Google DeepMind in February 2026, combining Pro-level image quality with Flash-level speed — generating a 4K image in just 4–6 seconds. Unlike traditional image models, Nano Banana 2 integrates directly into the standard Chat Completions API, with no separate image endpoint required. Just describe what you need in conversation to generate an image, and continue editing across turns — for example, generate a product shot first, then change the background to a sunset scene with a single sentence. It also supports real-time visual grounding from the web, accurately rendering specific landmarks, branded products, and other real-world objects. ### Free Agent & Reasoning Models Xiaomi's MiMo series is purpose-built for complex reasoning, function calling, and tool use — well-suited to autonomous agent workflows that require multi-step planning and chained tool execution. | Model | Highlights | | :---------------------------------------------------------------------------- | :----------------------------------------------------- | | [xiaomi-mimo-v2-pro-free](https://aihubmix.com/model/xiaomi-mimo-v2-pro-free) | Advanced reasoning, function calling, 1T+ params | | [xiaomi-mimo-v2.5-free](https://aihubmix.com/model/xiaomi-mimo-v2.5-free) | 1.02T params, 42B active, 1M context, 1000+ tool calls | ## Top 5 Free Models on AIHubMix 🔥 ### [coding-glm-5.1-free](https://aihubmix.com/model/coding-glm-5.1-free) — Best Free Coding Model Released by Zhipu AI in April 2026 with around 754B parameters. **GLM-5.1 became the first open-source model to top SWE-bench Pro at 58.4%** — surpassing GPT-5.4 (57.7%), Claude Opus 4.6 (57.3%), and Gemini 3.1 Pro (54.2%). Across 12 benchmarks covering reasoning, coding, agents, tool use, and browsing, it shows a balanced capability profile suited to demanding developer workflows. Via AIHubMix, it's a drop-in upgrade for any Cursor, Cline, Aider, or Claude Code setup at zero cost. ### [coding-glm-5-free](https://aihubmix.com/model/coding-glm-5-free) — Open-Source Code Powerhouse GLM-5.1's predecessor: a 745B-parameter MoE architecture (44B active), released February 2026. Scored 77.8% on SWE-bench Verified, achieving open-source state-of-the-art on agent coding leaderboards including Terminal Bench 2.0, with overall coding ability on par with Claude Opus 4.5. ### [gpt-4.1-free](https://aihubmix.com/model/gpt-4.1-free) `Hot` — Best Free 1M-Context Model > Context 1M · Latency 0.529s · Throughput 72 TPS · Free input and output OpenAI's next-generation flagship released April 2025. Comprehensively surpasses GPT-4o on coding and instruction following — 54.6% SWE-bench Verified, 87.4% IFEval. The 1M ultra-long context is uniquely suited to large-scale document analysis, codebase understanding, and complex agent workflows. The free version is hosted on Azure, offering fast response and high stability. ### [xiaomi-mimo-v2-pro-free](https://aihubmix.com/model/xiaomi-mimo-v2-pro-free) `New` — Best Free Agent Model > Context 256K · Latency 1.673s · Throughput 41 TPS · Free input and output Xiaomi's large reasoning model — MoE architecture with over 1T total parameters and roughly 42B active during inference. Ranked **8th on the global Intelligence Index** (2nd among Chinese models). Coding capability surpasses Claude Sonnet 4.6, and overall agent capability approaches Opus 4.6 — making it a strong pick for complex code generation and long-chain multi-tool workflows. ### [xiaomi-mimo-v2.5-free](https://aihubmix.com/model/xiaomi-mimo-v2.5-free) — Strongest Free Open Reasoning Model The current top of the MiMo series, with an Artificial Analysis Intelligence Index score of **54**. Built on a hybrid-attention MoE architecture (1.02T total / 42B active) with a **1M-token context window**. Improves comprehensively over V2-Pro on general agent capability, complex software engineering, and long-horizon tasks — supporting agent workflows with **1,000+ tool calls in a single session**. *** ## AIHubMix vs Openrouter Which Free AI API Should You Pick? If you've searched "free AI API," "OpenRouter alternative," or "free Claude API," you've likely seen a fragmented landscape. OpenRouter is the most-cited name in this category, but its free tier and AIHubMix's free tier solve fundamentally different problems — one optimizes for **breadth of open-source models**, the other for **access to frontier proprietary models without paying**. ### Where OpenRouter wins * **Open-source variety** — if your work centers on DeepSeek, Llama 3.3, Qwen, or fine-tuned community models, OpenRouter's catalog is broader. * **Random free-model routing** — the `openrouter/free` virtual model picks any available free open-source model, useful for cheap fallback chains. * **Long-standing brand recognition** in the indie OSS community. ### Where AIHubMix wins * **Free access to closed-source frontier models** — GPT-5.5, GPT-Image-2, Gemini 3, and Claude-class capability via GLM-5.1 are usable at \$0. OpenRouter's free tier deliberately excludes these. * **Native Claude Code integration** — AIHubMix exposes both `/v1/chat/completions` (OpenAI format) and `/v1/messages` (Anthropic format with `anthropic-beta` and `anthropic-version` header forwarding). Drop in via `ANTHROPIC_BASE_URL` with no proxy or translation layer. * **Image generation in the same gateway** — call GPT-Image-2 or Nano Banana 2 with the same API key you use for chat. * **Multi-provider failover per model** — when one upstream throttles or degrades, requests transparently re-route, raising your effective ceiling beyond what a single-upstream gateway delivers. * **Higher cumulative free quota** — daily caps spread across 27+ models, not a single 200-request bucket. **When to pick AIHubMix:** you want OpenAI/Anthropic/Google flagship models for free, a single OpenAI-compatible endpoint, and image generation in the same gateway. **When to pick OpenRouter:** you only need open-source models (Llama, DeepSeek, Qwen, Gemma) and prefer the broadest open-source catalog over frontier proprietary access. ## How to Get a Free AI Model API Key (3 Steps) The full flow for accessing free models via AIHubMix: 1. **Sign up** at [aihubmix.com](https://aihubmix.com) — email or OAuth, no credit card. 2. **Create an API key** on the **API Keys** page. Format: `sk-...` 3. **Pick a model** from the [free model catalog](https://aihubmix.com/models) and start calling. ## Use Cases & Integrations ### Free Models in Claude Code (Anthropic CLI) [Claude Code](https://docs.aihubmix.com/cn/api/Claude-Code) is Anthropic's official AI coding CLI, now a core part of many developer workflows. With a one-line environment variable, you can route Claude Code through AIHubMix and use **any free coding model as the backend** — no Anthropic billing required. ```bash theme={null} export ANTHROPIC_BASE_URL="https://aihubmix.com" export ANTHROPIC_AUTH_TOKEN="sk-YOUR_KEY" claude ``` Practical routing strategy: hand off everyday code generation to [kimi-for-coding-free](https://aihubmix.com/model/kimi-for-coding-free) or [coding-glm-5.1-free](https://aihubmix.com/model/coding-glm-5.1-free), use [gpt-4.1-free](https://aihubmix.com/model/gpt-4.1-free) for documentation and comments, and let [xiaomi-mimo-v2-pro-free](https://aihubmix.com/model/xiaomi-mimo-v2-pro-free) handle planning and orchestration of complex tasks. The full dev-assist pipeline runs at zero inference cost. See the [Claude Code integration docs](https://docs.aihubmix.com/cn/api/Claude-Code) for setup details — also available directly on [Claude Desktop](https://docs.aihubmix.com/cn/api/claude-desktop). ### Free Models in Cursor, Cline, Aider, and Other AI Coding Editors Any AI coding editor that supports a custom OpenAI-compatible endpoint works with AIHubMix free models. Configure `https://aihubmix.com/v1` as the base URL and pick a `*-free` model — drop-in replacement for paid GPT-5 or Claude usage in IDE assistants. ### Free Models in AI Agents & Autonomous Workflows [OpenClaw](https://docs.aihubmix.com/cn/api/OpenClaw) — open-source autonomous AI agent platform released November 2025, currently with **3.2M+ users**. Supports nearly every mainstream messaging channel — WhatsApp, Telegram, Slack, Discord — letting AI agents execute tasks directly inside platforms users already work in. Through AIHubMix, both [xiaomi-mimo-v2-pro-free](https://aihubmix.com/model/xiaomi-mimo-v2-pro-free) and [coding-glm-5.1-free](https://aihubmix.com/model/coding-glm-5.1-free) work seamlessly as backend models with full support for function calling, multi-turn context, and structured output. [Hermes Agent](https://docs.aihubmix.com/cn/api/hermes-agent-ai-hubmix) — NousResearch's agent framework, deeply optimized for tool use and structured JSON output. Its `execute_code` tool compresses multi-step pipelines into a single inference call, dramatically reducing round trips. Ideal for automation pipelines requiring strict JSON output — AIHubMix's automatic rate-limit rotation across providers ensures long-running tasks aren't interrupted when a single provider hits its quota. ### Free Models with Open-Source Clients AIHubMix is an officially supported API provider for several popular open-source applications: * **Desktop chat clients** — [Cherry Studio](https://docs.aihubmix.com/cn/clients/Cherry-Studio) is one of the most popular local AI chat clients, with a clean UI and convenient multi-model management. Select AIHubMix as the API provider to use GPT-4.1, Gemini Flash, GLM-5.1, and other free models in a desktop chat interface. * **Multi-model proxy & translation** — [LiteLLM](https://docs.aihubmix.com/cn/clients/LiteLLM) provides unified call management and load balancing across multiple free models; [NextAI Translator](https://docs.aihubmix.com/cn/clients/NextAITranslator) supports free models for high-quality multilingual translation. * **MCP / IDE integrations** — Claude Desktop, Continue, Open WebUI, and any tool that accepts an OpenAI-compatible endpoint. *** ## Rate Limits & Free Quota Free models on AIHubMix operate under per-model rate limits expressed as **requests per minute (RPM)** and **daily token caps**. Specifics are listed on each model's page at [aihubmix.com/models](https://aihubmix.com/models). Compared to single-provider free tiers: * **More headroom than OpenRouter** — multiple providers backing each model, with automatic failover when one upstream throttles. * **Higher cumulative ceiling than Google AI Studio** — instead of 1,500 req/day on a single model, AIHubMix lets you spread traffic across 27+ free models. * **No surprise expiry** — quotas reset daily; no 30-day trial cliff. For production traffic, the recommended pattern is **paid quota for the critical path, free models for auxiliary workloads** (batch summarization, log enrichment, draft generation, non-revenue-critical features). ## FAQ **Q: Why choose AIHubMix over OpenRouter, AIMLAPI, or Google AI Studio?** A: AIHubMix offers a unified OpenAI-compatible API aggregating **500+ global models including 27+ continuously updated free models** — and unlike OpenRouter, the free tier includes frontier proprietary models like GPT-5.5, GPT-Image-2, and Gemini 3 (not just open-source). Paid models are priced more competitively. The platform is officially operated by AIHubMix, LLC (USA) with formal authorization from major cloud vendors — making it trustworthy on both stability and compliance. **Q: Do I need a credit card to use AIHubMix free models?** A: No. Sign up with email or OAuth, create an API key, and start calling. Free models are usable immediately without any payment method on file. **Q: Do free models on AIHubMix have a time limit or trial expiry?** A: No trial expiry. Free models remain available within their respective per-minute and daily quotas indefinitely. Limits are expressed as RPM and daily token caps — see each model's page for specifics. **Q: Which free model offers the strongest overall coding capability?** A: As of May 2026, [coding-glm-5.1-free](https://aihubmix.com/model/coding-glm-5.1-free) leads — its 58.4% SWE-bench Pro score surpasses GPT-5.4 (57.7%), Claude Opus 4.6 (57.3%), and Gemini 3.1 Pro (54.2%), making it the **first open-source model to top the SWE-bench Pro leaderboard**. [kimi-for-coding-free](https://aihubmix.com/model/kimi-for-coding-free) particularly excels at multi-file context understanding and code refactoring. **Q: Are AIHubMix free models suitable for production?** A: For moderate production traffic, yes — with careful quota planning. AIHubMix's automatic failover balances load across multiple providers, increasing effective available quota. For higher-traffic production scenarios, run core inference on paid quota and route auxiliary work (batch summarization, log analysis, non-critical paths) to free models for a cost/stability balance. **Q: Can I use AIHubMix free models with the OpenAI Python or Node.js SDK?** A: Yes — AIHubMix is fully OpenAI-compatible. Set `base_url` to `https://aihubmix.com/v1` and use any official OpenAI SDK, LangChain integration, LlamaIndex pipeline, or AI gateway. No code rewrite required. **Q: Does AIHubMix support free image generation APIs?** A: Yes. Free image generation includes **GPT-Image-2** (OpenAI's first reasoning-capable image model, up to 4096×4096) and **Nano Banana 2** (`gemini-3.1-flash-image-preview-free`, 4K in 4–6 seconds). Both are accessed through standard chat-completions or image endpoints — no separate billing or quota system. *** ## Get Started Today Ready to ship AI features without burning your runway? Sign up at [aihubmix.com](https://aihubmix.com), grab a free API key, and start calling 27+ frontier models in minutes. For deeper integration guides, model performance specs, quota details, and SDK examples, see the [AIHubMix official documentation](https://docs.aihubmix.com/en). The complete free model catalog lives at [aihubmix.com/models](https://aihubmix.com/models). **Related guides:** [Claude Code setup](https://docs.aihubmix.com/cn/api/Claude-Code) · [Cherry Studio integration](https://docs.aihubmix.com/cn/clients/Cherry-Studio) · [LiteLLM gateway](https://docs.aihubmix.com/cn/clients/LiteLLM) · [OpenClaw agent platform](https://docs.aihubmix.com/cn/api/OpenClaw) · [Hermes Agent for structured output](https://docs.aihubmix.com/cn/api/hermes-agent-ai-hubmix) *** **References & Sources** * [Introducing GPT-4.1 | OpenAI](https://openai.com/index/gpt-4-1/) * [MiMo-V2-Pro | Xiaomi](https://mimo.xiaomi.com/mimo-v2-pro) * [MiMo-V2.5-Pro | Xiaomi](https://mimo.xiaomi.com/mimo-v2-5-pro/) * [GLM-5.1 | Hugging Face](https://huggingface.co/zai-org/GLM-5.1) * [GLM-5.1 Overview | Z.AI Developer Docs](https://docs.z.ai/guides/llm/glm-5.1) * [GLM-5.1 SWE-bench Pro Results | VentureBeat](https://venturebeat.com/technology/ai-joins-the-8-hour-work-day-as-glm-ships-5-1-open-source-llm-beating-opus-4) * [GLM Coding Plan | Zhipu AI](https://www.bigmodel.cn/glm-coding) * [OpenClaw | Official Docs](https://docs.openclaw.ai/) * [Hermes Agent | Nous Research](https://hermes-agent.nousresearch.com/) * [Claude Code LLM Gateway Docs | Anthropic](https://code.claude.com/docs/en/llm-gateway) *Last updated: May 7, 2026*