Model Intelligence Directory
Verified intelligence on publicly available AI models, capabilities, modalities, API availability, licensing, deployment options, and official documentation.
AI21 Labs
Jamba 1.5 is AI21 Labs’ hybrid SSM–Transformer model generation documented for long-context enterprise text workloads via AI21’s platform and published model cards.
Browse AI21 Labs models AI21 Labs profile Compare companies · AI21 Labs
Amazon
Amazon Nova Lite is a cost-efficient Nova family model on Amazon Bedrock documented for high-volume multimodal and text tasks.
Amazon
Amazon Nova Pro is part of Amazon’s Nova foundation model family available through Amazon Bedrock, with capabilities and modalities documented in AWS Bedrock model pages.
Browse Amazon models Amazon profile Compare companies · Amazon
Anthropic
Claude 3 Haiku is Anthropic’s speed/cost-oriented Claude 3 model for high-throughput and near-real-time applications.
Anthropic
Claude 3 Opus is the highest-capability model in Anthropic’s Claude 3 family announcement, intended for complex analysis and reasoning workloads.
Anthropic
Claude 3.5 Haiku is Anthropic’s speed-oriented Claude 3.5-generation model documented for near-real-time assistants and high-throughput workloads via Claude.ai and the Anthropic API.
Anthropic
Claude 3.5 Sonnet is a Claude 3.5 family model from Anthropic positioned for strong general performance across coding, analysis, and vision workloads, available via Claude.ai and the Anthropic…
Anthropic
Claude 3.7 Sonnet is an Anthropic Claude model generation documented for hybrid reasoning / extended thinking modes alongside general assistant workloads. Confirm current naming and IDs on Anthropic’s…
Anthropic
Claude Opus 4 is Anthropic’s highest-capability Claude 4-generation model as documented for complex reasoning and analysis workloads.
Anthropic
Claude Sonnet 4 is part of Anthropic’s Claude 4 model generation documented on Anthropic’s models pages for balanced speed and capability across coding and analysis tasks.
Browse Anthropic models Anthropic profile Compare companies · Anthropic
Black Forest Labs
FLUX.1 is a text-to-image model family from Black Forest Labs, offered in multiple tiers (e.g., [pro], [dev], [schnell]) with documentation on Black Forest Labs’ site and partner hosts.
Browse Black Forest Labs models Compare companies · Black Forest Labs
Cohere
Command R is Cohere’s retrieval-augmented generation oriented model family documented for enterprise RAG, tool use, and multilingual workflows.
Cohere
Command R+ is Cohere’s larger Command R family model optimized for retrieval-augmented generation, tool use, and multilingual enterprise workflows, documented in Cohere’s model docs.
Browse Cohere models Cohere profile Compare companies · Cohere
DeepSeek
DeepSeek-R1 is DeepSeek’s reasoning-oriented model line, with open materials describing reinforcement-learning-style training for reasoning behaviors.
DeepSeek
DeepSeek-V3 is a large mixture-of-experts language model released by DeepSeek with open technical reporting and weights distribution described on DeepSeek’s channels and Hugging Face.
Browse DeepSeek models DeepSeek profile Compare companies · DeepSeek
Gemini 1.5 Flash is a Gemini 1.5-generation model documented by Google for fast, cost-efficient multimodal workloads with long-context support as described in Google’s model docs.
Gemini 1.5 Pro is a Google DeepMind/Google multimodal model in the Gemini family, notable in Google’s documentation for long-context processing across modalities.
Gemini 2.0 Flash is a Gemini 2.0-generation model positioned by Google for fast, cost-efficient multimodal agents and app workloads.
Gemini 2.5 Flash is Google’s speed/cost-oriented Gemini 2.5 model documented for high-volume multimodal applications via Gemini API and Vertex AI.
Gemini 2.5 Pro is a Google Gemini generation model documented on Google AI / Gemini API model pages for advanced reasoning and multimodal application workloads.
Imagen 3 is Google’s text-to-image model generation available through Google’s generative AI products (e.g., Gemini / Vertex AI Image) as documented by Google.
Browse Google models Google profile Compare companies · Google
Meta
Llama 3.1 is Meta’s open-weight Llama generation announced with multiple parameter scales and an expanded context window relative to earlier Llama 3 releases, distributed under Meta’s Llama license…
Meta
Llama 3.2 expands the Llama family with lightweight text models and vision-capable models as described in Meta’s release materials.
Meta
Llama 3.3 is a Meta Llama text model release positioned as a high-capability instruct model in the Llama 3 line (see Meta’s publication for parameter scale and evaluations).
Mistral AI
Codestral is Mistral AI’s code-focused generative model documented for code completion and software engineering assistance via Mistral’s platform and related releases.
Mistral AI
Mistral Large is Mistral AI’s flagship text model offered via La Plateforme and partner clouds, with capabilities and versions described in Mistral’s documentation.
Mistral AI
Mistral Nemo is a Mistral model collaboration documented with NVIDIA, available as open weights and via Mistral’s platform depending on the published distribution channel.
Mistral AI
Mistral Small is a Mistral AI hosted model documented on Mistral’s models pages for cost-efficient multilingual text generation via La Plateforme.
Browse Mistral AI models Mistral AI profile Compare companies · Mistral AI
OpenAI
DALL·E 3 is OpenAI’s text-to-image system integrated into ChatGPT and the OpenAI Images API, with prompt following improvements described in OpenAI’s announcement.
OpenAI
GPT-4 Turbo refers to OpenAI’s GPT-4-class API models marketed for improved cost/latency characteristics relative to earlier GPT-4 API snapshots, with details defined per model ID in OpenAI’s model…
OpenAI
GPT-4.1 is an OpenAI GPT-4-class API model generation documented on OpenAI’s models pages for improved coding and instruction-following relative to earlier GPT-4o / GPT-4 Turbo snapshots. Exact limits…
OpenAI
GPT-4o is OpenAI’s flagship omni model family generation announced for ChatGPT and the OpenAI API, designed to accept combinations of text and image inputs (with audio capabilities described…
OpenAI
GPT-4o mini is OpenAI’s cost-efficient GPT-4o-family model documented for high-volume chat and light multimodal workloads via the OpenAI API and ChatGPT surfaces where offered.
Browse OpenAI models OpenAI profile Compare companies · OpenAI
Technology Innovation Institute
Falcon 180B is a large open language model released by the Technology Innovation Institute (TII) with weights and license terms published for research and commercial use under the…
Browse Technology Innovation Institute models Compare companies · Technology Innovation Institute
xAI
Grok-2 is xAI’s Grok model generation announced for the X platform and xAI API, with capabilities described in xAI’s announcements. Prefer xAI’s current docs for model IDs and…
xAI
Grok-3 is an xAI Grok model generation announced for xAI API and X product surfaces. Prefer xAI’s current documentation for model IDs, limits, and feature availability.
