Search tools, models…
Ctrl
K
Create
Ctrl
K
Build
AI Models
Foundation models for text, code, and multimodal work.
See our full best-of picks
39 tools
AI Models
AI Agents
Coding Tools
JavaScript Frameworks
No-code / Low-code
Website Builders
Backend & Databases
APIs
Project Management
ChatGPT
@chatgpt.com
OpenAI's assistant, the front door to GPT models for most people.
Claude
@claude.ai
Anthropic's assistant, known for writing quality and long documents.
Claude Fable 5
@platform.claude.com
Anthropic's most capable widely released model, for demanding reasoning and long-horizon agentic work.
Claude Haiku 4.5
@anthropic.com
Near-frontier quality at commodity prices and high speed.
Claude Mythos 5
@anthropic.com
Anthropic's limited-availability frontier model, Fable-tier pricing on a new tokenizer.
Claude Opus 4.8
@anthropic.com
Anthropic's deepest reasoner for the hardest problems.
Claude Opus 5
@anthropic.com
Anthropic's Opus-tier flagship: tops the Artificial Analysis Intelligence Index at roughly half of Claude Fable 5's price.
Claude Sonnet 5
@anthropic.com
Anthropic's workhorse: frontier coding and agents at mid-tier pricing.
DeepSeek V3.2
@deepseek.com
Open-weight frontier-class reasoning at a fraction of the cost.
DeepSeek V4 Flash
@deepseek.com
The cheapest model in the catalog scoring above 50, at $0.14 per million in.
DeepSeek V4 Pro
@deepseek.com
DeepSeek's flagship reasoning model: frontier-class quality at open-weight pricing.
Gemini
@gemini.google.com
Google's assistant, tightly integrated with Search and Workspace.
Gemini 2.5 Flash
@deepmind.google
Google's fast, cheap multimodal model for production workloads.
Gemini 3 Pro
@deepmind.google
Google's frontier multimodal model with a million-token context.
Gemini 3.7 Flash
@deepmind.google
381 tokens a second, the fastest frontier-class model Artificial Analysis measures.
Gemini Embedding 2
@deepmind.google
Five modalities natively, 100+ languages, and Matryoshka dimension truncation.
GLM-5.3
@z.ai
Z.ai's frontier release: 93 tokens a second and under two seconds to first token.
GPT-5.1
@openai.com
OpenAI's flagship with adaptive reasoning and a huge ecosystem.
GPT-5.6 Luna
@openai.com
OpenAI's cheap, fast 5.6 variant: 147 tokens a second at $0.20 per million in.
GPT-5.6 Sol
@openai.com
OpenAI's frontier flagship, strongest at command-line and multi-step coding tasks.
GPT-5.6 Terra
@openai.com
The mid tier of OpenAI's 5.6 line: most of Sol's quality at 40% of the price.
Grok 4.1
@x.ai
xAI's frontier model with native access to real-time X data.
Grok 4.5
@x.ai
xAI's fastest, most token-efficient frontier model yet, tuned on real coding-agent data.
Grok 4.6
@x.ai
Frontier-tier scores at $2 per million input tokens, the cheapest model at that level.
Hugging Face
@huggingface.co
The GitHub of machine learning: models, datasets, and demos.
Jina Embeddings v4
@jina.ai
2048 dimensions across text, image and PDF, with dimension truncation.
Kimi K2
@moonshot.ai
Moonshot's open-weight agentic model punching far above its price.
Kimi K3
@moonshot.ai
Moonshot's largest model yet, competitive with frontier closed models.
Llama 4 Maverick
@llama.com
Meta's open-weight multimodal MoE you can run yourself.
Meta Muse Spark 1.1
@ai.meta.com
Meta's first proprietary flagship model, succeeding the open-weight Llama line.
Meta Muse Spark 1.2
@ai.meta.com
Meta's current multimodal flagship, 57 on the Intelligence Index.
MiniMax M3
@minimax.io
Open-weight, 1M context, $0.30 per million input tokens.
Mistral Large 3
@mistral.ai
Europe's frontier lab, with strong models and open-weight options.
OpenAI text-embedding-3-large
@openai.com
The default choice for years, now behind the open-weight leaders on MTEB.
OpenRouter
@openrouter.ai
One API for every major LLM, with unified billing and routing.
Qwen3 235B
@qwen.ai
Alibaba's strongest open-weight family, excellent multilingual coverage.
Qwen3-Embedding-8B
@qwen.ai
Beats every closed API model on MTEB, and the weights are open.
Qwen3.8-Max
@qwen.ai
Alibaba's 2.4-trillion-parameter multimodal flagship, currently in preview.
Voyage Multimodal 3.5
@voyageai.com
Compact 1024-dimension vectors, built for storage-constrained indexes.
