Model Catalog

OpenAI's newest frontier reasoning model for the most complex coding and professional work, with text and image input and configurable reasoning effort.

Highest-compute variant of GPT-5.5 that thinks harder to deliver consistently better answers on the hardest professional tasks.

Affordable frontier reasoning model for complex coding and professional tasks with text and image input.

OpenAI's strongest mini model for coding, computer use, and subagents at lower cost.

OpenAI's cheapest GPT-5-class reasoning model for simple, high-volume tasks with text and image input.

Higher-compute variant of GPT-5.4 for more reliable answers on demanding professional tasks.

Reasoning model for realtime voice interactions with stronger instruction following and reliable tool use for voice-agent workflows.

Default realtime voice model for natural two-way audio conversations, voice agents, and customer support.

Streaming speech-to-speech translation model for live multilingual audio, priced per minute of audio.

Streaming speech-to-text model for low-latency realtime transcription, priced per minute of audio.

Speech-to-text model powered by GPT-4o with improved word error rate and language recognition over original Whisper.

Lighter, lower-cost speech-to-text model for high-volume transcription.

State-of-the-art image generation and editing model accepting text and high-fidelity image inputs.

Video generation model priced per second of generated video.

Higher-quality video generation model priced per second of generated video.

OpenAI's most capable embedding model for English and non-English text retrieval and similarity tasks.

Cost-efficient embedding model for text retrieval and similarity at high volume.

Anthropic's most capable model for complex reasoning, long-horizon agentic coding, and high-autonomy work, available on Bedrock.

Anthropic's balanced model offering the best combination of speed and intelligence with extended and adaptive thinking.

Anthropic's fastest model with near-frontier intelligence and extended thinking support.

Opus 4.5 reasoning model with extended thinking, still available on Bedrock.

Balanced Sonnet 4.5 model with extended thinking, available on Bedrock.

Fast, cost-efficient Claude 3.5 generation model still listed as available on Bedrock.

Claude Opus 4.1, deprecated by Anthropic and scheduled for retirement on August 5, 2026.

Amazon's balanced multimodal model offering strong accuracy, speed, and cost across text, image, and video inputs.

Amazon's cost-efficient next-gen multimodal model for automation, document processing, and customer support across text, images, and video.

Amazon's low-cost multimodal model that processes text, image, and video inputs for tasks like document analysis and visual Q&A.

Amazon's fastest text-only model, optimized for speed and low cost in summarization, translation, and classification.

Amazon's image generation model that creates studio-quality images from text and image prompts with watermarking and content moderation.

Amazon's video generation model that creates short videos from text and image prompts with camera motion controls.

Amazon's speech-to-speech model enabling natural, real-time voice conversations with low latency and multilingual support.

Amazon's second-generation text embeddings model with configurable output dimensions and improved retrieval accuracy.

Amazon's embedding model that converts text, images, and video into vector representations for search and retrieval.

Amazon's multimodal Nova model for complex reasoning and model distillation; marked Legacy with end-of-life September 14, 2026.

Meta's 17-billion active-parameter mixture-of-experts model with 128 experts, optimized for multimodal chat and instruction following.

Meta's 17-billion active-parameter mixture-of-experts model with 16 experts and a 10M-token context window for long-document tasks.

Meta's 70-billion-parameter model with improved efficiency, delivering strong reasoning and coding with a 128K context window.

Meta's 90-billion-parameter multimodal model that processes text and images for visual reasoning and document understanding.

Meta's 11-billion-parameter multimodal model that processes both text and images with a 128K context window.

Meta's largest open model with 405 billion parameters and a 128K context window, supporting tool use and multilingual tasks.

Meta's 70-billion-parameter model with a 128K context window and support for tool use and code generation.

Meta's compact 8-billion-parameter model with a 128K context window, suitable for edge deployment and fine-tuning.

Mistral AI's 675-billion-parameter model with strong performance on coding, reasoning, and multilingual tasks.

Mistral AI's flagship model with strong reasoning, multilingual support, and a 32K context window for complex enterprise tasks.

Mistral AI's 124-billion-parameter multimodal model that processes text and images for visual reasoning and document understanding.

Mistral AI's cost-efficient model optimized for low-latency tasks like classification, translation, and customer support.

Mistral AI's sparse mixture-of-experts model with 8 experts of 7B parameters each, delivering strong performance at fast inference speeds.

Cohere's model for complex RAG workflows, multi-step tool use, and enterprise tasks with a 128K context window.

Cohere's scalable LLM optimized for retrieval-augmented generation and tool use in enterprise applications with a 128K context window.

Cohere's unified multimodal embedding model that processes text, images, and mixed content in a single model for search and RAG.

OpenAI's open-weight 120-billion-parameter model available on Amazon Bedrock for reasoning and coding workloads.

OpenAI's open-weight 20-billion-parameter model available on Amazon Bedrock for lightweight reasoning and chat.

Google's most advanced reasoning model for complex problem-solving, deep reasoning, and frontier coding tasks.

Most intelligent flash-class model for sustained frontier performance at high volume.

Frontier-class flash model delivering performance rivaling larger models at lower cost.

Most cost-efficient multimodal model in the Gemini 3 family for high-throughput tasks.

High-quality, low-latency Live API model for real-time bidirectional voice and video dialogue.

Advanced reasoning and coding model with deep thinking capabilities for complex tasks.

Best price-performance model for low-latency, high-volume tasks with thinking support.

Fastest and most budget-friendly multimodal model for high-frequency, cost-sensitive tasks.

Native audio dialog model producing natural conversational speech from multimodal input.

Specialized model that controls a browser or UI by reasoning over screenshots to take actions.

High-quality Gemini-native image generation and editing model for detailed visual creation.

Gemini-native conversational image generation and editing model known as Nano Banana.

Google's dedicated text-to-image generation model offering Fast, Standard, and Ultra quality tiers.

State-of-the-art cinematic text-to-video and image-to-video generation model with native audio.

High-quality text-to-video and image-to-video generation model with synchronized audio.

Latest multimodal embedding model generating vectors from text, image, audio, and video inputs.

Text embedding model producing high-dimensional vectors for retrieval and similarity tasks.