Model Finder

Most popular local LLM - runs on consumer hardware with excellent performance

Anthropic's most intelligent model - excels at complex reasoning, coding, and analysis

OpenAI's flagship omni model - text, vision, and audio in one

Anthropic's most intelligent model - best in world for coding, agents, and computer use. Features effort parameter for controlling reasoning depth, zoom computer use, and extended thinking.

Best open speech recognition - 99 languages supported

Excellent balance of capability and efficiency for local deployment

Industry-standard image generation model with excellent quality and flexibility

OpenAI flagship video generation model. Creates realistic and imaginative videos up to 1 minute from text prompts. Understands motion, physics, and cinematic techniques.

Latest Llama with improved tool use and multilingual capabilities

Meta's largest and most capable open model with state-of-the-art performance

The original Mistral - set new standards for 7B models

Google's flagship with 2M token context - can process entire codebases

Small, fast, and cheap GPT-4o for most tasks

Best model in the world for agents, coding, and computer use. Enhanced domain knowledge in coding, finance, and cybersecurity.

State-of-the-art coding model - rivals GPT-4 on coding benchmarks

Classic SD model with massive ecosystem of fine-tunes and LoRAs

State-of-the-art image generation with exceptional quality

OpenAI's latest image generation model

Runway Gen-3 Alpha - major leap in video generation fidelity, consistency, and motion. Supports text-to-video, image-to-video, and motion brush controls.

Best-in-class English embedding model for RAG applications

8x faster Whisper with minimal quality loss

Industry-leading voice synthesis with cloning

Highly efficient MoE model - GPT-3.5 quality at fraction of compute

State-of-the-art MoE model rivaling GPT-4 at fraction of compute

Google's latest multimodal model with native tool use and real-time streaming

Hybrid reasoning model with superior performance for real-world coding and agentic tasks. Handles complex multi-step problems with rigor and attention to detail.

Multilingual large language model optimized for dialogue. Outperforms many open-source and closed models on common benchmarks. Supports 8 languages with 128K context.

Best coding model for consumer GPUs

Fast FLUX variant - 4-step generation with Apache 2.0 license

Google Veo 2 - generates 4K videos with cinematic quality. Understands real-world physics and human movement. Supports various styles from photorealistic to surreal.