Compare 90 AI models across benchmarks. Explore performance metrics, capabilities, and find the right model for your use case.
Claude 3 Haiku is Anthropic’s smallest and fastest Claude 3 model, optimized for low-latency, high-throughput text and lightweight vision tasks. It ta...
Claude 3 Opus is Anthropic’s flagship Claude 3 series large language model optimized for high-level reasoning, complex analysis, and long-context unde...
Claude 3 Sonnet is Anthropic's balanced Claude 3-series model that targets a strong mix of intelligence, speed, and cost-efficiency for general-purpos...
Claude 3.5 Haiku is Anthropic’s lightweight Claude 3.5-series model optimized for speed and low cost while retaining strong reasoning and coding capab...
Claude 3.5 Opus is Anthropic’s flagship Claude 3.5 series model optimized for high‑end reasoning, coding, and complex analysis. It offers strong perfo...
Claude.ai is Anthropic’s chat-oriented interface for accessing Claude family large language models via the web. It provides natural language assistanc...
Anthropic's Claude 4 flagship model with advanced reasoning...
Balanced Claude 4 model for everyday tasks...
Enhanced Claude 4 with improved agentic capabilities...
Anthropic's most capable model with extended thinking...
Claude 4.5 Sonnet with breakthrough coding performance...
Codestral is Mistral AI’s open-source, 22B-parameter code-specialized language model optimized for software development workflows such as code generat...
Cohere Rerank is a production search and retrieval reranking model that scores and reorders candidate documents based on their relevance to a query. I...
Cohere Command R+ is a production-grade large language model optimized for enterprise workloads, retrieval-augmented generation (RAG), and tool use. I...
Updated Command R+ with improved performance...
DeepSeek's reasoning model trained with pure RL...
Updated DeepSeek R1 with improved math reasoning...
State-of-the-art open-source MoE model with 671B parameters...
DeepSeek V3.1 with improved capabilities...
DeepSeek V3.2 experimental release...
Baidu's ERNIE 5.0 preview...
Zhipu's GLM-4.5 bilingual model...
Zhipu's latest GLM model...
The ChatGPT API is OpenAI's hosted interface to its GPT-4.1-class language models, exposed as a general-purpose text-in/text-out service. It supports ...
GPT-4.1 is a flagship OpenAI large language model that offers GPT-4-level intelligence with improved speed, cost, and reliability. It is designed for ...
OpenAI GPT-4.1 with 1M context window...
OpenAI's GPT-4.5 preview model...
OpenAI's GPT-5 with unified reasoning capabilities...
GPT-5 with maximum reasoning effort...
Updated GPT-5 with improved capabilities...
Gemini 2.0 Flash with extended reasoning capabilities...
xAI's Grok 4 with breakthrough reasoning capabilities...
Updated Grok 4 with enhanced thinking...
Grok 4.1 with extended reasoning mode...
xAI's flagship model with real-time knowledge...
Efficient version of Grok-2...
Moonshot's Kimi K2 with advanced reasoning...
Kimi K2 with extended thinking capabilities...
Meta Llama 3 is Meta’s third-generation open large language model family, released in 8B and 70B parameter sizes. It is optimized for instruction foll...
Llama 3.1 70B is a large-scale open-weight language model from Meta designed to provide near frontier-level performance in reasoning, coding, and gene...
Meta's efficient 70B model matching 405B performance...
Meta's Llama 4 Maverick with 1M context...
Mistral 7B is a dense 7B-parameter open-weight language model and Mixtral 8x7B is a sparse Mixture-of-Experts model with eight 7B experts, both from M...
Mistral FT refers to fine-tuned variants of Mistral AI base language models, exposed via the Mistral API for domain- or task-specific use. These model...
Mistral's flagship model with 123B parameters...
Mistral's flagship 2025 model...
Mistral Small is a lightweight proprietary instruction-tuned language model from Mistral AI, designed to offer strong reasoning and coding performance...
Mistral 7B and Mixtral 8x22B are open-weight large language models from Mistral AI, designed for efficient, high-quality text generation and reasoning...
Mixtral 8x7B is a sparse mixture-of-experts large language model by Mistral AI, combining eight 7B expert networks with conditional routing for high e...
NVIDIA's open model optimized for inference...
OpenAI's reasoning model with extended thinking capabilities...
Smaller, cost-efficient reasoning model optimized for coding...
Preview version of OpenAI's o1 reasoning model...
Alibaba's flagship open model with strong multilingual support...
Specialized coding model rivaling GPT-4o on code tasks...
Alibaba's Qwen 3 flagship model...
01.AI's ultra-fast inference model...
OpenAI's o3 advanced reasoning model...
Efficient reasoning model optimized for speed...
Amazon's fast, cost-effective multimodal model...
Amazon's balanced multimodal model...
GPT-4o ("omni") is OpenAI's flagship multimodal model that natively supports text, vision, and audio. It is optimized for fast, low-latency interactio...
GPT-4o mini is a lightweight, cost-efficient member of the GPT-4o family optimized for fast, low-latency text and vision tasks. It supports multimodal...
Gemini API is Google’s unified interface for accessing Gemini family multimodal models that can understand and generate text, code, and images, and re...
Gemini 1.5 Flash is Google's lightweight, high-throughput multimodal model in the Gemini 1.5 family, optimized for low latency and cost while still su...
Gemini 1.5 Pro is Google’s flagship multimodal large language model capable of understanding and generating text, code, and analyzing images, audio, a...
Google's fastest multimodal model with native tool use...
Google's Gemini 2.5 Pro with native reasoning...
Compact multimodal Llama for edge deployment...
Meta's first multimodal Llama with vision capabilities...
Meta's Llama 4 Scout with 10M context and vision...
Mistral's first multimodal model...
Mistral's flagship multimodal model...
GPT-4o Vision is OpenAI's multimodal GPT-4o variant optimized for understanding images and returning text outputs. It can interpret documents, UI scre...
Google Gemini Vision is the multimodal vision component of Google's Gemini family, designed to interpret and reason over images and other visual input...
Midjourney is a proprietary text-to-image generative model accessed primarily via a Discord bot and web interface. It specializes in producing high-qu...
Stable Diffusion XL (SDXL) is a high-capacity latent diffusion text-to-image model by Stability AI designed for photorealistic and artistic image gene...
Jina Embeddings are a family of text embedding models developed by Jina AI, optimized for semantic search, retrieval, and other vector-based NLP tasks...
Voyage Law Embeddings is a domain-specialized text embedding model optimized for legal documents, case law, contracts, and regulatory text. It is desi...