Browse all AI Gateway models
Every model available on Vercel AI Gateway, with API access, pricing, and a playground. 413 models · Page 1 of 7.
Search and filter all models →- CallstackApexApex is Callstack’s agentic coding model for mobile and web development, optimized for React Native, React, and Next.js. With a 256K context window, it supports feature development, debugging, code reviews, testing, and large refactors.
- QuiverAIArrow 2SVG generation model with agentic refinement and tool calling.
- QuiverAIArrow 2 TelosHigher-refinement Arrow 2 variant combining Arrow's speed with frontier-model reasoning.
- ReflectionBeam 1.0Beam is Reflection’s reasoning model for coding and agentic workflows, combining strong coding capabilities with fast, low-latency inference.
- BrowserbaseBrowserbase FetchFetch the contents of a URL with Browserbase, optionally through proxies, returning the upstream response body unchanged or content extracted as Markdown or JSON. Use it as a built-in tool through the AI Gateway to let any model read a page it already has a URL for.
- BrowserbaseBrowserbase SearchSearch the web with Browserbase's Search API for fast, structured web discovery, returning result titles, URLs, and metadata without opening a browser session. Use it as a built-in tool through the AI Gateway to give any model access to real-time web information.
- ByteDanceBytedance Seed 1.8Bytedance Seed 1.8 features stronger multimodal understanding and agent capabilities. The model delivers superior performance across a wide range of complex real-world tasks, helping enterprises create greater value.
- ScaleDownClassify-1ScaleDown Classify-1 is a task-specific small language model for text classification. It assigns text to custom labels and returns probability scores, supporting intent detection, content categorization, and routing workflows. Its deterministic, encoder-only approach produces consistent classification results without sampling variability.
- AnthropicClaude Fable 5Claude Fable 5 is a Mythos-class model with robust safeguards. It can handle long-running, complex, and asynchronous tasks where previous models would have needed more frequent check-ins.
- AnthropicClaude Fable 5.1Claude Fable 5.1 improves on Fable 5 across long-running agentic coding, knowledge work, and research, following instructions precisely over sessions that run unattended for hours.
- AnthropicClaude Haiku 4.5Claude Haiku 4.5 matches Sonnet 4's performance on coding, computer use, and agent tasks at substantially lower cost and faster speeds. It delivers near-frontier performance and Claude’s unique character at a price point that works for scaled sub-agent deployments, free tier products, and intelligence-sensitive applications with budget constraints.
- AnthropicClaude Haiku 5.5Claude Haiku 5.5 brings adaptive reasoning to Anthropic’s low-cost Haiku family for everyday tasks, tool use, and scaled sub-agent deployments. It is the first Haiku with adjustable effort levels, letting developers balance reasoning depth and token usage for each task, with the option to turn thinking off for simpler requests.
- AnthropicClaude Opus 4.5Claude Opus 4.5 is Anthropic’s latest model in the Opus series, meant for demanding reasoning tasks and complex problem solving. This model has improvements in general intelligence and vision compared to previous iterations. In addition, it is suited for difficult coding tasks and agentic workflows, especially those with computer use and tool use, and can effectively handle context usage and external memory files.
- AnthropicClaude Opus 4.6Opus 4.6 is the world’s best model for coding and professional work, built to power agents that take on whole categories of real-world work. It excels across the entire SDLC, breaking through on hard problems, identifying complex bugs, and demonstrating deeper codebase understanding. It also delivers a step-change in knowledge work, with near-production-ready documents, presentations, and spreadsheets on the first pass.
- AnthropicClaude Opus 4.7Opus 4.7 builds on the coding and agentic strengths of Opus 4.6 with stronger performance on complex, multi-step tasks and more reliable agentic execution. It also brings improved performance on knowledge work, from drafting documents to building presentations and analyzing data.
- AnthropicClaude Opus 4.8Opus 4.8 is a focused upgrade to Opus 4.7 and is Anthropic's best generally available model for coding, agentic tasks, and enterprise workflows. It builds on the strengths of previous Opus models with stronger performance on complex, multi-step coding tasks. Anthropic recommends using it on long-horizon coding and agentic tasks. It is also stronger on professional work, including document drafting, data analysis, and presentations.
- AnthropicClaude Opus 5Claude Opus 5 is the latest model in Anthropic's Opus family and a step-change improvement over Opus 4.8. It delivers major gains over Opus 4.8 in agentic coding, professional knowledge work, and long-horizon reasoning, and it is stronger per token across effort levels.
- AnthropicClaude Opus 5.5Claude Opus 5.5 is a step-change improvement over Opus 5 for agentic coding, long-running agentic tasks, knowledge work, communication, and vision. It is more token-efficient per completed task and reports progress, findings, and next steps in clear language.
- AnthropicClaude Sonnet 4.5Claude Sonnet 4.5 is the newest model in the Sonnet series, offering improvements and updates over Sonnet 4.
- AnthropicClaude Sonnet 4.6Claude Sonnet 4.6 is the most capable Sonnet-class model yet, with frontier performance across coding, agents, and professional work. It excels at iterative development, complex codebase navigation, end-to-end project management with memory, polished document creation, and confident computer use for web QA and workflow automation.
- AnthropicClaude Sonnet 5Sonnet 5 is an upgrade to Sonnet 4.6, with gains across agentic coding and professional work. It builds on the strengths of previous Sonnet models, bringing top-tier intelligence at Sonnet pricing for coding, agents, and everyday professional work at scale.
- AnthropicClaude Sonnet 5.5Claude Sonnet 5.5 is a step-change improvement over Sonnet 5 for well-scoped everyday work like building features, fixing bugs, and creating polished documents, slides, and spreadsheets. It also writes more clearly than Sonnet 5.
- MistralCodestral EmbedCode embedding model that can embed code databases and repositories to power coding assistants.
- CohereCohere Rerank 3.5A model that allows for re-ranking English Language documents and semi-structured data (JSON).
- CohereCohere Rerank 4 FastA light version of Rerank 4 Pro, this is a multilingual model that allows for re-ranking English and non-english documents and semi-structured data (JSON). This model is better suited for low latency and high throughput use-cases than its pro variant.
- CohereCohere Rerank 4 ProA multilingual model that allows for re-ranking English and non-english documents and semi-structured data (JSON). This model is better suited for state-of-the-art quality and complex use-cases than its fast variant.
- CohereCohere Rerank 5 FastCohere Rerank 5 Fast is a multilingual text reranking model optimized for low latency and high throughput in search and agent workflows. It ranks documents and semi-structured data across more than 100 languages, with calibrated relevance scores and a 32,768-token context window per query-document pair.
- CohereCohere Rerank 5 ProCohere Rerank 5 Pro is a multilingual text reranking model for complex queries and long or domain-specific documents. It ranks documents and semi-structured data across more than 100 languages, with calibrated relevance scores and a 32,768-token context window per query-document pair.
- CohereCommand ACommand A is Cohere's most performant model to date, excelling at tool use, agents, retrieval augmented generation (RAG), and multilingual use cases. Command A has a context length of 256K, only requires two GPUs to run, and has 150% higher throughput compared to Command R+ 08-2024.
- PerplexityDecider Flash V1.2 4BPerplexity Decider Flash V1.2 is a 4B multimodal decision model for high-volume classification, routing, labeling, and filtering. It evaluates text, structured JSON, and images against yes/no questions, custom choices, and ordered rubrics, returning typed answers with probabilities. It supports a 256K context window.
- PerplexityDecider V1.2 27BPerplexity Decider V1.2 is a 27B multimodal decision model for classification, routing, and scoring. It evaluates text, structured JSON, and images against yes/no questions, custom choices, and ordered rubrics, returning typed answers with probabilities. It supports multiple questions over shared content and a 256K context window.
- Microsoft AIDecision-1Microsoft Decision-1 is a text decision model from Microsoft AI for classification, routing, and evaluation. It answers yes/no questions, selects among predefined options, and scores inputs against custom criteria, returning calibrated probabilities and confidence values. With support for multiple questions per request and a 32,768-token context window, it helps developers triage support requests, guide automated workflows, and evaluate AI responses against a rubric.
- DeepSeekDeepSeek R1 0528The DeepSeek R1 model has undergone a minor version upgrade, with the current version being DeepSeek-R1-0528.
- DeepSeekDeepSeek V3.1DeepSeek-V3.1 is post-trained on the top of DeepSeek-V3.1-Base, which is built upon the original V3 base checkpoint through a two-phase long context extension approach, following the methodology outlined in the original DeepSeek-V3 report. We have expanded our dataset by collecting additional long documents and substantially extending both training phases. The 32K extension phase has been increased 10-fold to 630B tokens, while the 128K extension phase has been extended by 3.3x to 209B tokens. Additionally, DeepSeek-V3.1 is trained using the UE8M0 FP8 scale data format to ensure compatibility with microscaling data formats.
- DeepSeekDeepSeek V3.2DeepSeek-V3.2: Official successor to V3.2-Exp.
- DeepSeekDeepSeek V3.2 Thinking
- DeepSeekDeepseek V4 FlashDeepSeek-V4 series incorporate several key upgrades in architecture and optimization: (1) a hybrid attention architecture that combines Compressed Sparse Attention (CSA) and Heavily Compressed Attention (HCA) to improve long-context efficiency; (2) ManifoldConstrained Hyper-Connections (mHC) that enhance conventional residual connections; (3) and the Muon optimizer for faster convergence and greater training stability
- DeepSeekDeepSeek V4 Flash 0731DeepSeek-V4 series incorporate several key upgrades in architecture and optimization: (1) a hybrid attention architecture that combines Compressed Sparse Attention (CSA) and Heavily Compressed Attention (HCA) to improve long-context efficiency; (2) ManifoldConstrained Hyper-Connections (mHC) that enhance conventional residual connections; (3) and the Muon optimizer for faster convergence and greater training stability
- DeepSeekDeepSeek V4 Flash Vision ExpDeepSeek V4 Flash Vision Exp is an experimental vision-enabled version of DeepSeek V4 Flash, adding image understanding while retaining its text, agent, reasoning, and world-knowledge capabilities. It is suited for document and chart understanding, visual question answering, and multimodal agent workflows.
- DeepSeekDeepSeek V4 ProDeepSeek-V4 series incorporate several key upgrades in architecture and optimization: (1) a hybrid attention architecture that combines Compressed Sparse Attention (CSA) and Heavily Compressed Attention (HCA) to improve long-context efficiency; (2) ManifoldConstrained Hyper-Connections (mHC) that enhance conventional residual connections; (3) and the Muon optimizer for faster convergence and greater training stability
- DeepSeekDeepSeek V4 Pro 0813This is the 8/13 updated weights version of DeepSeek V4 Pro.
- DeepSeekDeepSeek V4.1 FlashDeepSeek V4.1 Flash is an AI model from DeepSeek designed for fast, efficient multimodal interactions. It combines native multimodal capabilities with a new architecture focused on stronger performance, faster responses, and lower computational cost.
- DeepSeekDeepSeek V4.1 Flash 101710/17 snapshot of DeepSeek V4.1 Flash
- DeepSeekDeepSeek V4.1 Flash 1017 FastHigher-throughput variant of DeepSeek V4.1 Flash 1017 served by Baseten for multimodal reasoning, tool calling, and structured output, with automatic prefix caching.
- DeepSeekDeepSeek V4.1 Flash FastHigher-throughput variant of DeepSeek V4.1 Flash for responsive coding, reasoning, and agentic workloads.
- Fish AudioDrama 3Drama 3 is Fish Audio’s text-to-speech model, currently in preview, that lets you direct voice performances using natural-language instructions. It supports control over emotion, pacing, pauses, and vocal sounds, with tags that apply effects to specific words. Multi-speaker dialogue and directions written in different languages make it well suited to expressive voiceovers, character dialogue, and audio storytelling.
- BasetenElephant-1Elephant-1 is a 9B decision model from Baseten's Base Labs for classification, routing, and scoring. It evaluates structured context against choice, yes/no, and ordered scoring questions and returns answer probabilities and confidence in a single forward pass, without generating free-form text. Applications can use these probabilities to select answers or set confidence thresholds.
- CohereEmbed 5 FastCohere's lightweight embedding model optimized for low latency and high throughput in interactive search, agent loops, and high-volume retrieval. Embed 5 Fast supports text, images, and mixed content, over 100 languages, and a 128,000-token context window, with output dimensions from 256 to 2,048. It shares an embedding space with Embed 5 Pro.
- CohereEmbed 5 ProCohere's most capable embedding model for quality-critical enterprise retrieval across text, images, and mixed content such as PDFs. Embed 5 Pro supports over 100 languages and a 128,000-token context window, with output dimensions from 256 to 2,048. It shares an embedding space with Embed 5 Fast, enabling Pro indexing with Fast querying.
- PerplexityEmbed v1 0.6bpplx-embed-v1 is a state-of-the-art text embedding model optimized for real-world, web-scale retrieval tasks.
- PerplexityEmbed v1 4bpplx-embed-v1 is a state-of-the-art text embedding model optimized for real-world, web-scale retrieval tasks.
- CohereEmbed v4.0A model that allows for text, images, or mixed content to be classified or turned into embeddings.
- FireworksEmber-1Ember-1 is Fireworks’ research-preview reasoning model built on Kimi K3 for coding and agentic workflows. It supports image input, function calling, and a 1M-token context window. Fireworks reports approximately 40% fewer generated tokens while maintaining comparable quality across its evaluations, helping reduce the cost of workloads with extensive reasoning.
- ExaExa SearchSearch the web with Exa's neural search API, returning LLM-optimized, ranked results with content excerpts. Use it as a built-in tool through the AI Gateway to give any model access to real-time web information.
- ScaleDownExtract-1ScaleDown Extract-1 is a task-specific language model for extracting structured data from unstructured text using a custom JSON schema. It identifies domain-specific entities with confidence scores, character offsets, and surrounding context, and supports nested objects and arrays for workflows such as invoice processing, contract analysis, and record parsing.
- Black Forest LabsFlux 3FLUX 3 is a new multimodal foundation model that learns from images, videos, and audio within a unified architecture.
- Black Forest LabsFLUX 3 ImageFLUX 3 Image is Black Forest Labs’ image generation and editing model. It supports text-to-image generation, editing with up to 10 reference images, and layouts specified with bounding boxes, across five resolution classes from 768 × 768 to 4K.
- Black Forest LabsFLUX 3 Image KleinFLUX 3 Image Klein is Black Forest Labs’ smaller, faster image generation and editing model, with 5.6 billion parameters and open weights. It supports text-to-image generation, image editing, multiple reference images, and layouts specified with bounding boxes. The hosted API produces images across five resolution classes from 768 × 768 to 4K.
- ProdiaFlux SchnellLightning-fast image generation
- Black Forest LabsFLUX.1 Fill [pro]A state-of-the-art inpainting model, enabling editing and expansion of real and generated images given a text description and a binary mask. This provider gives the option to change the moderation level for inputs and outputs. The control is under safety tolerance and is by default 2 on a range from 0 (more strict) through 6 (more permissive).