📚 AI Model Reviews
LLM, Image, Video, Coding. Comprehensive research index covering features and evaluations of major AI models.
Last Update: Feb 1, 2026
🧠 LLMs (Large Language Models)
GPT 5
"Strict but knowledgeable professor." Usually leads the industry with overwhelming reasoning ability and knowledge base.
Claude Opus 4.6
"The Orchestra Conductor". Evolved into a "Fortress of Intellect" with 1M tokens and Agent Teams.
Claude Sonnet 4.5
"Competent practitioner/editor." Perfect balance of speed and precision, smartly handling any task.
Gemini 3 Pro
"Multitasking genius engineer." The strongest practical partner, fully integrated with the Google ecosystem.
Gemini 3 Flash
"Light-speed messenger." Overwhelming speed and affordability. The best choice for API use and massive text processing.
Grok 3
"Galaxy-roaming roast comedian." A maverick with real-time X search and powerful reasoning capabilities.
DeepSeek V3
"Highly efficient mathematician." Open source yet approaching the pinnacle, a revolutionary in cost performance.
Llama 4
"Sturdy pioneer." The strongest model that runs locally. A free spirit loved by developers worldwide.
Mistral Large 2
"Sophisticated European researcher." Elegant model design pursuing extreme efficiency and precision.
Qwen 3
"Universal interpreter/mathematician." World-class multilingual processing and mathematical reasoning.
Gemma 3
"Compact researcher." Lightweight yet high-performance. Intelligence that runs on personal devices.
Kimi k2.5
"Librarian with long-term memory." Remembers entire books with a context window exceeding 2 million tokens.
Microsoft Copilot
"Expert office secretary." Dramatically improves business productivity by integrating with Office products.
🎨 Image Generation Models
Midjourney v7
"Solitary genius painter." Overwhelming artistry and aesthetics, the pinnacle of AI art.
Flux 2
"Precise architect." A monumental open model that changed the world with text rendering and fidelity. Commercial use OK.
Stable Diffusion 3.5
"Sturdy engineer." Freedom to run on home PCs and customizability. A fortress of open source resisting censorship.
Ideogram v3
"Tasteful graphic designer." Changes the norms of logo and poster production with the world's best text rendering capability.
Adobe Firefly
"Competent agency creative." Copyright cleared and commercial use OK. Accelerates practical work with Photoshop integration.
Imagen 4
"Realistic photographer." Perfect for business with realism mistaken for photos and accurate text rendering.
GPT Image 1.5
"Knowledgeable illustrator." Image generation that perfectly understands context born from dialogue with ChatGPT.
Grok Image
"Rebellious punk rocker." Real-time X integration and Flux-based descriptive power. A maverick swaying between regulation and freedom.
Nano Banana Pro
"Pop creator." Fast and controllable generation capability, a rising star supported by the SNS generation.
💻 AI Coding Agents
Cursor
"Crazy mad scientist." AI-native IDE that makes VS Code a thing of the past. Excels at overall design with Composer.
Windsurf
"Cool surfer moving at light speed." Explosive development with "Flow State" that never stops thinking and Cascade engine.
OpenAI Codex (App)
"The Abstract Architect." Manages a fleet of autonomous agents without writing a single line of code.
Antigravity
"Clumsy but versatile butler from the future." Google's secret weapon that even autonomously operates browsers. A pioneer of Vibe Coding.
GitHub Copilot
"Reliable veteran pair programmer." Robust security and overwhelming integration. When in doubt, choose this.
🎥 Video Generation Models
Sora 2
"Argumentative film director." A "World Simulator" where sound and vision merge. The long-awaited next-gen video generation model.
Veo 3
"Serious physicist." A laboratory that perfectly simulates light and fluid. Powerful integration with Gemini.
Runway Gen-4
"Avant-garde video artist." Overwhelming control with Motion Brush. An unlimited canvas for creators.
Kling AI 2.6
"Taciturn artisan." World-class character portrayal and lip-sync. A realist pursing overwhelming reality.
Midjourney Video
"Whimsical painter." The beauty of still images brought to video. Artistic visual beauty that ignores laws of physics.
🎼 Audio Generation Models
ElevenLabs v4
"Voice actor with a thousand masks." The definitive voice AI taking the world by storm with overwhelming emotional expression and multilingual support.
Suno v5
"Enthusiastic music producer." Finally reaching "broadcast level" audio quality. An era where anyone can be a hitmaker.
Udio v3
"Solitary conductor." Precise partial correction with Inpainting. For those who want to "construct" music rather than just "generate" it.
🧊 3D Generation Models
Rodin Gen-2
"Seeker of beauty." Overwhelming texture quality and PBR materials. The pinnacle of photorealistic 3D generation.
Meshy 4
"Master of structure." Strong in hard surfaces, a practical tool producing clean topology directly usable in games.
Hunyuan3D 2.5
"Sturdy blacksmith." Strongest shape reproduction in open source. Freedom to run locally and precise geometry.
Luma Genie 3
"Wizard in your pocket." AR production completed on a smartphone. Instantly shape ideas with intuitive operation.
Tripo 3D
"Rapid modeler." Generate high-quality 3D models in just seconds. Rigging support makes it instantly ready for apps.