API Status Page

This version of Qwen 2.5 adds vision capabilities to the solid foundation of its text-only sibling. It's great for multimodal tasks like analyzing images with text, understanding screenshots, or processing documents with visual elements. A practical choice when you need both language and vision understanding without breaking the bank.

This is Alibaba's crown jewel for complex reasoning. With 235B parameters but only 22B active per token, it delivers exceptional performance on math, science, and coding challenges. It's designed specifically for tasks that require deep, step-by-step thinking - think of it as your personal research assistant that never gets tired.

When you need raw capability without the reasoning overhead, this is your go-to. It's incredibly knowledgeable across domains and excels at tasks requiring broad understanding. Great for content creation, analysis, and general problem-solving where you don't need to see the thought process.

Purpose-built for developers, this model excels at everything from quick scripts to complex system design. It understands modern development practices, can work across multiple files, and even helps with debugging. If you're building software, this is like having a senior developer as your pair programming partner.

Takes the solid foundation of DeepSeek V3 and adds months of additional training and refinement. Better at following instructions, more knowledgeable, and still maintains that direct, no-nonsense communication style that made the original popular.

V3.1 is a capable general-purpose model that excels at coding, tool use, and complex problem-solving. It delivers strong performance across a wide range of tasks while maintaining good speed and efficiency.

When you enable thinking mode, V3.1 becomes a different beast entirely. It excels at complex mathematical problems, intricate coding challenges, and multi-step reasoning tasks. The thinking process is thorough and often educational.

Terminus takes everything learned from V3.1 and adds stability improvements. Better at maintaining conversation context, more reliable on long tasks, and improved tool-use capabilities. It's the version you'd deploy in production systems.

Terminus takes everything learned from V3.1 and adds stability improvements. Better at maintaining conversation context, more reliable on long tasks, and improved tool-use capabilities. It's the version you'd deploy in production systems.

V3.2 introduces DeepSeek Sparse Attention, making it incredibly efficient for long-context tasks while maintaining top-tier performance. It's designed for the modern AI workflow - handling massive codebases, long documents, and complex multi-step processes with ease.

Built from the ground up with reasoning in mind, this model excels at tasks that would challenge even human experts. It achieved gold medals in international math and programming competitions - that's how good it is at systematic problem-solving.

DeepSeek R1 proved that open-source reasoning models could compete with the best closed models. Whether you choose the original, the distilled versions, or the latest updates, you're getting genuine reasoning capabilities that show their work and think deeply about problems.

DeepSeek R1 proved that open-source reasoning models could compete with the best closed models. Whether you choose the original, the distilled versions, or the latest updates, you're getting genuine reasoning capabilities that show their work and think deeply about problems.

Similar to the Llama distilled model, but distilled on Qwen 32B instead. Slightly better at code, slightly more likely to fall into thought loops.

Flash 2.0 is like the Swiss Army knife of AI models - fast, reliable, and handles almost anything you throw at it. Its standout feature is the enormous context window, letting you work with entire codebases or long documents without breaking a sweat.

Takes everything great about Flash 2.0 and adds improved capabilities. Better at complex tasks while maintaining that signature Google speed. Perfect for when you need quick but thoughtful responses.

Combines Flash's speed with reasoning capabilities, letting you see the thought process behind answers. Great for educational use, debugging, or building trust in AI-assisted decisions.

Gemini 2.5 Flash-Lite is a member of the Gemini 2.5 series of models, a suite of highly-capable, natively multimodal models. Gemini 2.5 Flash-Lite is Google’s most cost-efficient model, striking a balance between efficiency and quality.

Gemini 2.5 Flash-Lite is a member of the Gemini 2.5 series of models, a suite of highly-capable, natively multimodal models. Gemini 2.5 Flash-Lite is Google’s most cost-efficient model, striking a balance between efficiency and quality. This version has "thinking" capabilities that enable it to provide responses with greater accuracy and nuanced context handling.

Gemini 2.5 Flash Image Preview is a state of the art image generation model with contextual understanding. It is capable of image generation, edits, and multi-turn conversations.

Similar to 2.0 Flash, but even faster. Not as smart, but still good at most things.

When you need Google's most capable model, 2.5 Pro delivers. It excels at mathematical reasoning, scientific analysis, and complex coding challenges. The thinking capabilities make it particularly valuable for research and development work.

Google's Imagen 4 is a powerful image generation model that creates high-quality, photorealistic images from text prompts. Built on advanced diffusion techniques and trained on diverse datasets. 2 images per prompt.

Google's Imagen 4 Ultra is a powerful image generation model that creates high-quality, photorealistic images from text prompts. Built on advanced diffusion techniques and trained on diverse datasets. 1 image per prompt.

Gemini 3 Pro represents Google's previous advances in AI. It's exceptionally good at complex reasoning tasks, long-context understanding, and maintaining coherence over extended conversations. Think of it as Google's answer to the most challenging cognitive tasks.

Gemini 3.1 Pro Preview is Google's frontier reasoning model, delivering enhanced software engineering performance, improved agentic reliability, and more efficient token usage across complex workflows. Building on the multimodal foundation of the Gemini 3 series, it combines high-precision reasoning across text, image, video, audio, and code with a 1M-token context window.

Nano Banana Pro is a state of the art image generation model with contextual understanding. It is capable of image generation, edits, and multi-turn conversations.

The fastest model in Google's Gemini 3 family, but don't confuse speed with simplicity. It handles complex tasks remarkably well while maintaining low latency. Perfect for real-time applications and rapid prototyping.

Proves that fast doesn't mean shallow. This model can quickly work through complex problems while showing its reasoning process. Ideal for interactive applications where users need both speed and transparency.

Scout brings vision capabilities to the Llama family while maintaining efficiency. It's designed for applications that need to understand both text and images without the computational overhead of larger models.

Maverick excels at maintaining natural, coherent conversations across long contexts. It's particularly good at tasks requiring nuanced understanding and can handle both text and visual inputs with impressive capability.

MiniMax M2 redefines what's possible with efficient AI design. Despite 'only' activating 10B parameters from its 230B total, it delivers performance that rivals much larger models. It's specifically optimized for coding workflows, making it perfect for development tools.

MiniMax-M2.1 is a lightweight, state-of-the-art large language model optimized for coding, agentic workflows, and modern application development. With only 10 billion activated parameters, it delivers a major jump in real-world capability while maintaining exceptional latency, scalability, and cost efficiency. Compared to its predecessor, M2.1 delivers cleaner, more concise outputs and faster perceived response times. It shows leading multilingual coding performance across major systems and application languages.

MiniMax-M2.5 is a high-efficiency large language model optimized for end-to-end coding and productivity. Building on MiniMax-M2, it extends strong coding and tool-use performance into broader office tasks, while handling multi-software context switching and cross-team collaboration across human and agent workflows. It reports strong benchmark performance, including 80.2% on SWE-Bench Verified, 51.3% on Multi-SWE-Bench, and 76.3% on BrowseComp, while improving token efficiency through planning-focused training.

Kimi K2 proved that Chinese AI companies could build world-class open models. It's particularly strong on coding tasks, mathematical problem-solving, and workflows. The 1 trillion parameter count sounds impressive, but the real magic is in how efficiently it uses them.

The September update brings improved capabilities and extended context length. Better at handling long documents, maintaining conversation coherence, and executing complex multi-step tasks.

When you enable thinking mode, K2 becomes a different model entirely. It excels at systematic problem-solving, showing detailed reasoning steps, and tackling challenges that require deep analysis.

Kimi K2.5 is Moonshot AI's native multimodal model, delivering state-of-the-art visual coding capability. Built on Kimi K2 with continued pretraining over approximately 15T mixed visual and text tokens, it delivers strong performance in general reasoning, visual coding, and agentic tool-calling.

Kimi K2.5 is Moonshot AI's native multimodal model, delivering state-of-the-art visual coding capability. Built on Kimi K2 with continued pretraining over approximately 15T mixed visual and text tokens, it delivers strong performance in general reasoning, visual coding, and agentic tool-calling.

A medium-sized open-weight model from OpenAI suitable for general-purpose tasks. gpt-oss-20b is an open-weight 21B parameter model released by OpenAI under the Apache 2.0 license. It uses a Mixture-of-Experts (MoE) architecture with 3.6B active parameters per forward pass, optimized for lower-latency inference and deployability on consumer or single-GPU hardware. The model is trained in OpenAI's Harmony response format and supports reasoning level configuration, fine-tuning, and capabilities including function calling, tool use, and structured outputs.

A large open-weight model from OpenAI. gpt-oss-120b is an open-weight, 117B-parameter Mixture-of-Experts (MoE) language model from OpenAI designed for high-reasoning and general-purpose production use cases. It activates 5.1B parameters per forward pass and is optimized to run on a single H100 GPU with native MXFP4 quantization. The model supports configurable reasoning depth, full chain-of-thought access, and native tool use, including function calling, browsing, and structured output generation.

Like gpt-4o, but faster. This model sacrifices some of the original GPT-4o's precision for significantly reduced latency. It accepts both text and image inputs.

GPT-4.1 brings significant improvements in code generation, instruction following, and complex reasoning. It's particularly good at software development tasks and maintains coherence over long contexts.

GPT-4.1 Mini is a mid-sized model delivering performance competitive with GPT-4o at substantially lower latency. It has a very large context window and scores 45.1% on hard instruction evals, 35.8% on MultiChallenge, and 84.1% on IFEval. Mini also shows strong coding ability (e.g., 31.6% on Aider's polyglot diff benchmark) and vision understanding.

For tasks that demand low latency, GPT‑4.1 nano is the fastest model in the GPT-4.1 series. It delivers exceptional performance at a small size with its 1 million token context window, and scores 80.1% on MMLU, 50.3% on GPQA, and 9.8% on Aider polyglot coding – even higher than GPT‑4o mini. It's ideal for tasks like classification or autocompletion.

GPT-5 represents a significant leap in AI capability. It excels at professional tasks, complex problem-solving, and maintaining natural conversation. The different variants (Instant, Thinking, Pro) let you choose the right tool for your specific needs.

OpenAI's latest flagship model. PhD-level intelligence at most things. This version has reasoning capabilities, and is not well suitied to general chat, but is great for complex reasoning tasks.

A lighter-weight GPT-5 variant optimized for speed while retaining strong reasoning and tool use.

An ultra-fast GPT-5 variant tuned for low-latency tasks with reasoning and tool use.

Building on GPT-5's foundation, 5.1 brings enhanced capabilities and better performance on complex tasks. The different variants let you choose the right tool for your specific needs.

Building on GPT-5's foundation, 5.1 brings enhanced reasoning capabilities and better performance on complex tasks. The reasoning variants are particularly impressive for analytical work.

GPT-5.2 achieves something special - it's both faster and smarter than its predecessors. It excels at specialized knowledge work while maintaining conversational warmth. The speed improvements make it practical for real-time applications.

GPT-5.2 achieves something special - it's both faster and smarter than its predecessors. It excels at specialized knowledge work while maintaining conversational warmth. The speed improvements make it practical for real-time applications.

GPT-5.2 achieves something special - it's both faster and smarter than its predecessors. It excels at specialized knowledge work while maintaining conversational warmth. The speed improvements make it practical for real-time applications.

The o3 family represents OpenAI's focus on systematic reasoning. These models excel at mathematical problems, scientific analysis, and multi-step reasoning tasks. They're designed for when you need to think through problems methodically.

Proves that you don't need massive size for sophisticated reasoning. o4-mini delivers impressive analytical capabilities while maintaining reasonable speed and cost. Perfect for applications that need reasoning power without the overhead.

The o3 family represents OpenAI's focus on systematic reasoning. These models excel at mathematical problems, scientific analysis, and multi-step reasoning tasks. They're designed for when you need to think through problems methodically.

When you encounter problems that require deep, systematic analysis, o3 Pro is your best bet. It uses additional compute time to work through complex challenges methodically, often achieving results that surprise even AI researchers.

OpenAI's previous image generation model, using lots of crazy tech like custom tools for text and reflections. This model generates 1 image per prompt.

OpenAI's latest and greatest image generation model, using lots of crazy tech like custom tools for text and reflections. This model generates 1 image per prompt.

GLM-4.5 is an open-weight MoE model that competes with o3 and Claude 4 while being smaller and stronger than DeepSeek-R1 and Kimi K2. It excels at coding and is trained using the Muon architecture, the same one to train Kimi K2.

GLM-4.5 is an open-weight MoE model that competes with o3 and Claude 4 while being smaller and stronger than DeepSeek-R1 and Kimi K2. It excels at reasoning and coding and is trained using the Muon architecture, the same one to train Kimi K2. This variant has reasoning mode enabled for step-by-step thinking.

GLM-4.5V is an open-weight MoE model that competes with o3 and Claude 4 while being smaller and stronger than DeepSeek-R1 and Kimi K2. It excels at coding and is trained using the Muon architecture, the same one to train Kimi K2.

GLM-4.5V is an open-weight MoE model that competes with o3 and Claude 4 while being smaller and stronger than DeepSeek-R1 and Kimi K2. It excels at reasoning and coding and is trained using the Muon architecture, the same one to train Kimi K2. This variant has reasoning mode enabled for step-by-step thinking.

GLM-4.5-Air is the lightweight variant of GLM-4.5, an open-weight MoE model that competes with o3 and Claude 4 while being smaller and stronger than DeepSeek-R1 and Kimi K2. It excels at coding and is trained using the Muon architecture, the same one to train Kimi K2.

GLM-4.5 Air is the lightweight variant of GLM-4.5, an open-weight MoE model that competes with o3 and Claude 4 while being smaller and stronger than DeepSeek-R1 and Kimi K2. It excels at reasoning and coding and is trained using the Muon architecture, the same one to train Kimi K2. This variant has reasoning mode enabled for step-by-step thinking.

Compared with GLM-4.5, this generation brings several key improvements - Longer context window: The context window has been expanded from 128K to 200K tokens, enabling the model to handle more complex tasks. Superior coding performance: The model achieves higher scores on code benchmarks and demonstrates better real-world performance in applications such as Claude Code、Cline、Roo Code and Kilo Code, including improvements in generating visually polished front-end pages. Advanced problem-solving: GLM-4.6 shows a clear improvement in analytical performance and supports tool use during inference, leading to stronger overall capability. Enhanced tool use: GLM-4.6 exhibits stronger performance in tool using and search-based workflows, and integrates more effectively within frameworks. Refined writing: Better aligns with human preferences in style and readability, and performs more naturally in role-playing scenarios.

Compared with GLM-4.5, this generation brings several key improvements - Longer context window: The context window has been expanded from 128K to 200K tokens, enabling the model to handle more complex tasks. Superior coding performance: The model achieves higher scores on code benchmarks and demonstrates better real-world performance in applications such as Claude Code、Cline、Roo Code and Kilo Code, including improvements in generating visually polished front-end pages. Advanced reasoning: GLM-4.6 shows a clear improvement in reasoning performance and supports tool use during inference, leading to stronger overall capability. Enhanced tool use: GLM-4.6 exhibits stronger performance in tool using and search-based workflows, and integrates more effectively within frameworks. Refined writing: Better aligns with human preferences in style and readability, and performs more naturally in role-playing scenarios.

GLM-4.6V is a large multimodal model designed for high-fidelity visual understanding and long-context reasoning across images, documents, and mixed media. It supports up to 128K tokens, processes complex page layouts and charts directly as visual inputs, and integrates native multimodal function calling to connect perception with downstream tool execution. The model also enables interleaved image-text generation and UI reconstruction workflows, including screenshot-to-HTML synthesis and iterative visual editing.

GLM-4.7 is Z.AI's latest flagship model, featuring upgrades in two key areas: enhanced programming capabilities and more stable multi-step execution. It demonstrates significant improvements in executing complex tasks while delivering more natural conversational experiences and superior front-end aesthetics.

GLM-4.7 is Z.AI's latest flagship model, featuring upgrades in two key areas: enhanced programming capabilities and more stable multi-step reasoning/execution. It demonstrates significant improvements in executing complex tasks while delivering more natural conversational experiences and superior front-end aesthetics. This variant has reasoning mode enabled for step-by-step thinking.

GLM-5 is Z.ai’s flagship open-source foundation model engineered for complex systems design and long-horizon agent workflows. Built for expert developers, it delivers production-grade performance on large-scale programming tasks, rivaling leading closed-source models. With advanced agentic planning, deep backend reasoning, and iterative self-correction, GLM-5 moves beyond code generation to full-system construction and autonomous execution.

GLM-5 is Z.ai’s flagship open-source foundation model engineered for complex systems design and long-horizon agent workflows. Built for expert developers, it delivers production-grade performance on large-scale programming tasks, rivaling leading closed-source models. With advanced agentic planning, deep backend reasoning, and iterative self-correction, GLM-5 moves beyond code generation to full-system construction and autonomous execution.

xAI's flagship model that breaks records on lots of benchmarks (allegedly). Possesses deep domain knowledge in finance, healthcare, law, and science.

xAI's flagship model that excels at data extraction, coding, and text summarization. Possesses deep domain knowledge in finance, healthcare, law, and science.

A lightweight model that thinks before responding. Great for simple or logic-based tasks that do not require deep domain knowledge.

Grok 4 Fast is xAI's latest multimodal model with SOTA cost-efficiency and a 2M token context window.

Grok 4 Fast is xAI's latest multimodal model with SOTA cost-efficiency and a 2M token context window.

Grok 4.1 Fast is xAI's latest multimodal model with SOTA cost-efficiency and a 2M token context window.

Grok 4.1 Fast is xAI's latest multimodal model with SOTA cost-efficiency and a 2M token context window.