AI Models & Tools Comparison 2026

Cursor

A code editor with AI built in that can write whole features while you watch.

This one: Multi-model IDE, Tab autocomplete, 8 parallel subagents, BugBot PR reviewer

GitHub Copilot

Autocomplete on steroids inside your editor — it finishes your code as you type.

This one: Deep GitHub integration, largest user base, autonomous coding agent on issues

Kiro (AWS)

An AI pair-programmer that lives in your code editor and writes, explains, and fixes code with you.

This one: Spec-driven development workflow, agent hooks, deep AWS integration

Windsurf

An AI pair-programmer that lives in your code editor and writes, explains, and fixes code with you.

This one: Cascade agentic mode, proprietary SWE-1 models, auto lint fixing

Replit

An AI pair-programmer that lives in your code editor and writes, explains, and fixes code with you.

This one: Full cloud IDE with hosting/deployment, no local setup, Design Mode

Bolt.new

An AI pair-programmer that lives in your code editor and writes, explains, and fixes code with you.

This one: Entire dev environment in browser via WebContainers, 1M+ sites deployed

v0 (Vercel)

An AI pair-programmer that lives in your code editor and writes, explains, and fixes code with you.

This one: Image-to-code, one-click Vercel deployment, Next.js ecosystem

Lovable

An AI pair-programmer that lives in your code editor and writes, explains, and fixes code with you.

This one: Figma-like visual editor, one-click Supabase, $300M ARR

Claude Code

An AI pair-programmer that lives in your code editor and writes, explains, and fixes code with you.

This one: Terminal-native, deep agentic capabilities, extended thinking, no GUI overhead

Devin

An AI pair-programmer that lives in your code editor and writes, explains, and fixes code with you.

This one: Most autonomous tool, parallel agents, end-to-end development

Augment Code

An AI pair-programmer that lives in your code editor and writes, explains, and fixes code with you.

This one: Context Engine for large codebases, 200K context window, Memories

Tabnine

An AI pair-programmer that lives in your code editor and writes, explains, and fixes code with you.

This one: Privacy-first: air-gapped/on-prem, BYOM, 600+ languages

Amazon Q Developer

An AI pair-programmer that lives in your code editor and writes, explains, and fixes code with you.

This one: Deep AWS integration, automated Java 8-to-17 code transforms

JetBrains AI CZ

An AI pair-programmer that lives in your code editor and writes, explains, and fixes code with you.

This one: Deepest IDE integration, leverages JetBrains code analysis engine

Sourcegraph Cody/Amp

An AI pair-programmer that lives in your code editor and writes, explains, and fixes code with you.

This one: Code search across massive monorepos, unconstrained token usage (Amp)

VS Code + Copilot

An AI pair-programmer that lives in your code editor and writes, explains, and fixes code with you.

This one: Most-used IDE worldwide, multi-model AI agent, open-source extensions ecosystem

Xcode 26

An AI pair-programmer that lives in your code editor and writes, explains, and fixes code with you.

This one: On-device Swift AI, Claude/Codex agents, Apple ecosystem integration

Trae (ByteDance)

An AI pair-programmer that lives in your code editor and writes, explains, and fixes code with you.

This one: 100% free Cursor alternative, Claude/GPT/Gemini/DeepSeek, Builder no-code mode

Google Antigravity

An AI pair-programmer that lives in your code editor and writes, explains, and fixes code with you.

This one: Google's agentic IDE, Gemini 3-powered agents, cloud dev environment, autonomous coding

OpenCode

An AI pair-programmer that lives in your code editor and writes, explains, and fixes code with you.

This one: Open-source Claude Code alternative, 120K+ GitHub stars, TUI with Bubble Tea, LSP integration

Cline

An AI pair-programmer that lives in your code editor and writes, explains, and fixes code with you.

This one: Agentic coding in VSCode, terminal access, open-source

Continue

An AI pair-programmer that lives in your code editor and writes, explains, and fixes code with you.

This one: Open-source, multi-model support, local-first IDE plugin

Aider

An AI pair-programmer that lives in your code editor and writes, explains, and fixes code with you.

This one: Git-native terminal AI coding, multi-model, top SWE-Bench scores

Zed AI

An AI pair-programmer that lives in your code editor and writes, explains, and fixes code with you.

This one: Rust-based editor, ultra-fast performance, multi-model AI built-in

ChatGPT / GPT-5

A smart assistant you chat with — ask it to write, code, explain, summarize, or plan, and it answers in seconds.

This one: Flagship frontier model, massive ecosystem, native image gen

GPT-5.4 / GPT-5.4 Pro

A smart assistant you chat with — ask it to write, code, explain, summarize, or plan, and it answers in seconds.

This one: Native computer-use, 1M context, SWE-Bench Pro 57.7%, ARC-AGI-2 73.3%

Claude (Anthropic)

A smart assistant you chat with — ask it to write, code, explain, summarize, or plan, and it answers in seconds.

This one: Extended thinking, 1M token context, strong safety alignment

Gemini (Google)

A smart assistant you chat with — ask it to write, code, explain, summarize, or plan, and it answers in seconds.

This one: Deep Google Workspace integration, competitive Flash pricing

Gemini 3.1 Flash-Lite

A smart assistant you chat with — ask it to write, code, explain, summarize, or plan, and it answers in seconds.

This one: 2.5x faster than Gemini 2.5 Flash, cheapest in 3.x family, 64K output tokens

Llama (Meta)

A smart assistant you chat with — ask it to write, code, explain, summarize, or plan, and it answers in seconds.

This one: Open-weight, 10M token context (Scout), self-hostable

Mistral

A smart assistant you chat with — ask it to write, code, explain, summarize, or plan, and it answers in seconds.

This one: European AI sovereignty, mix of open + proprietary models

Grok (xAI)

A smart assistant you chat with — ask it to write, code, explain, summarize, or plan, and it answers in seconds.

This one: Real-time X/Twitter data, 2M token context

Perplexity AI

A smart assistant you chat with — ask it to write, code, explain, summarize, or plan, and it answers in seconds.

This one: Real-time web search with cited sources, aggregates best models

DeepSeek

A powerful, free AI chatbot and coder with no daily limits.

This one: 95% cheaper than GPT-4 Turbo, fully open source (MIT), o1-level reasoning

Qwen 3.5 (Alibaba)

A smart assistant you chat with — ask it to write, code, explain, summarize, or plan, and it answers in seconds.

This one: Largest open-source MoE family, 1M context, 201 languages, Apache 2.0

Mercury 2 (Inception)

A smart assistant you chat with — ask it to write, code, explain, summarize, or plan, and it answers in seconds.

This one: First diffusion LLM, 1000+ tok/s (5x faster), 1.7s end-to-end, 128K context

ERNIE 5.0 / Wenxin (Baidu)

A smart assistant you chat with — ask it to write, code, explain, summarize, or plan, and it answers in seconds.

This one: 2.4T parameters, native omni-modal, Baidu ecosystem integration

Doubao 2.0 (ByteDance)

A smart assistant you chat with — ask it to write, code, explain, summarize, or plan, and it answers in seconds.

This one: 90% cheaper than GPT-5.2, 200M+ users, TikTok/ByteDance ecosystem

Kimi K2.5 (Moonshot)

A smart assistant you chat with — ask it to write, code, explain, summarize, or plan, and it answers in seconds.

This one: Agent Swarm (100 parallel sub-agents), 1T params, image-to-code

GLM-5 (Zhipu AI)

A smart assistant you chat with — ask it to write, code, explain, summarize, or plan, and it answers in seconds.

This one: 2M context (largest Chinese), first Chinese AI IPO, CogView/CogVideo

Hunyuan 2.0 / Yuanbao (Tencent)

A smart assistant you chat with — ask it to write, code, explain, summarize, or plan, and it answers in seconds.

This one: Hybrid Transformer-Mamba arch, full creative ecosystem, WeChat integration

MiniMax M2.5

A smart assistant you chat with — ask it to write, code, explain, summarize, or plan, and it answers in seconds.

This one: Broadest multimodal (text+img+video+voice+music), rivals Opus at 1/10 price

Baichuan M3

A smart assistant you chat with — ask it to write, code, explain, summarize, or plan, and it answers in seconds.

This one: World's leading medical AI, clinical reasoning, runs on consumer GPUs

MiniCPM-V 4.5 (Tsinghua)

A smart assistant you chat with — ask it to write, code, explain, summarize, or plan, and it answers in seconds.

This one: 8B multimodal on mobile/edge, Apache 2.0, rivals GPT-4V at 1/100 size

Yi (01.AI)

A smart assistant you chat with — ask it to write, code, explain, summarize, or plan, and it answers in seconds.

This one: Strong bilingual EN/ZH LLM, open weights, competitive benchmarks

Command R+ (Cohere)

A smart assistant you chat with — ask it to write, code, explain, summarize, or plan, and it answers in seconds.

This one: Enterprise RAG specialist, 128K context, strong tool use

Phi-4 (Microsoft)

A smart assistant you chat with — ask it to write, code, explain, summarize, or plan, and it answers in seconds.

This one: 14B params beating much larger models on reasoning benchmarks

Reka

A smart assistant you chat with — ask it to write, code, explain, summarize, or plan, and it answers in seconds.

This one: Natively multimodal — processes text, image, video, audio together

GPT-5.4

A smart assistant you chat with — ask it to write, code, explain, summarize, or plan, and it answers in seconds.

This one: 1M ctx · OpenAI

Grok 4.20

A smart assistant you chat with — ask it to write, code, explain, summarize, or plan, and it answers in seconds.

This one: 2M ctx · xAI

Claude Opus 4.7

A smart assistant you chat with — ask it to write, code, explain, summarize, or plan, and it answers in seconds.

This one: 1M ctx · Anthropic

GPT-5.5

A smart assistant you chat with — ask it to write, code, explain, summarize, or plan, and it answers in seconds.

This one: 1M ctx · OpenAI

Grok 4.3

A smart assistant you chat with — ask it to write, code, explain, summarize, or plan, and it answers in seconds.

This one: 1M ctx · xAI

GPT-5.5 Instant

A smart assistant you chat with — ask it to write, code, explain, summarize, or plan, and it answers in seconds.

This one: OpenAI

Gemini 3.5 Flash

A smart assistant you chat with — ask it to write, code, explain, summarize, or plan, and it answers in seconds.

This one: 1M ctx · Google DeepMind

Claude Opus 4.8

A smart assistant you chat with — ask it to write, code, explain, summarize, or plan, and it answers in seconds.

This one: 1M ctx · Anthropic

Claude Fable 5

A smart assistant you chat with — ask it to write, code, explain, summarize, or plan, and it answers in seconds.

This one: 1M ctx · Anthropic

Grok Build 0.1

A smart assistant you chat with — ask it to write, code, explain, summarize, or plan, and it answers in seconds.

This one: 256K ctx · xAI

DeepSeek-V4-Pro

A smart assistant you chat with — ask it to write, code, explain, summarize, or plan, and it answers in seconds.

This one: 1M ctx · DeepSeek

DeepSeek-V4-Flash

A smart assistant you chat with — ask it to write, code, explain, summarize, or plan, and it answers in seconds.

This one: 1M ctx · DeepSeek

Qwen3.5-Omni

A smart assistant you chat with — ask it to write, code, explain, summarize, or plan, and it answers in seconds.

This one: 262K ctx · Alibaba

Qwen3.6-Plus

A smart assistant you chat with — ask it to write, code, explain, summarize, or plan, and it answers in seconds.

This one: 1M ctx · Alibaba

Qwen3.6-35B-A3B

A smart assistant you chat with — ask it to write, code, explain, summarize, or plan, and it answers in seconds.

This one: 262K ctx · Alibaba

Qwen3.6-Max-Preview

A smart assistant you chat with — ask it to write, code, explain, summarize, or plan, and it answers in seconds.

This one: 262K ctx · Alibaba

Qwen3.7-Max

A smart assistant you chat with — ask it to write, code, explain, summarize, or plan, and it answers in seconds.

This one: 1M ctx · Alibaba

GLM-5.1

A smart assistant you chat with — ask it to write, code, explain, summarize, or plan, and it answers in seconds.

This one: 200K ctx · Zhipu AI

GLM-5.2

A smart assistant you chat with — ask it to write, code, explain, summarize, or plan, and it answers in seconds.

This one: 1M ctx · Zhipu AI

GLM-5-Turbo

A smart assistant you chat with — ask it to write, code, explain, summarize, or plan, and it answers in seconds.

This one: 200K ctx · Zhipu AI

GLM-5V-Turbo

A smart assistant you chat with — ask it to write, code, explain, summarize, or plan, and it answers in seconds.

This one: 203K ctx · Zhipu AI

Kimi K2.6

A smart assistant you chat with — ask it to write, code, explain, summarize, or plan, and it answers in seconds.

This one: 262K ctx · Moonshot AI

Kimi K2.7 Code

A smart assistant you chat with — ask it to write, code, explain, summarize, or plan, and it answers in seconds.

This one: 262K ctx · Moonshot AI

MiniMax-M2.7

A smart assistant you chat with — ask it to write, code, explain, summarize, or plan, and it answers in seconds.

This one: 205K ctx · MiniMax

MiniMax-M3

A smart assistant you chat with — ask it to write, code, explain, summarize, or plan, and it answers in seconds.

This one: 1M ctx · MiniMax

ERNIE 5.1

A smart assistant you chat with — ask it to write, code, explain, summarize, or plan, and it answers in seconds.

This one: 128K ctx · Baidu

Hunyuan Hy3 Preview

A smart assistant you chat with — ask it to write, code, explain, summarize, or plan, and it answers in seconds.

This one: 256K ctx · Tencent

Doubao-Seed-2.0-lite

A smart assistant you chat with — ask it to write, code, explain, summarize, or plan, and it answers in seconds.

This one: 256K ctx · ByteDance

Step 3.7 Flash

A smart assistant you chat with — ask it to write, code, explain, summarize, or plan, and it answers in seconds.

This one: 256K ctx · StepFun

SenseNova U1

A smart assistant you chat with — ask it to write, code, explain, summarize, or plan, and it answers in seconds.

This one: SenseTime

Muse Spark

A smart assistant you chat with — ask it to write, code, explain, summarize, or plan, and it answers in seconds.

This one: Meta

Mistral Small 4

A smart assistant you chat with — ask it to write, code, explain, summarize, or plan, and it answers in seconds.

This one: 256K ctx · Mistral AI

Voxtral TTS

A smart assistant you chat with — ask it to write, code, explain, summarize, or plan, and it answers in seconds.

This one: Mistral AI

Mistral Medium 3.5

A smart assistant you chat with — ask it to write, code, explain, summarize, or plan, and it answers in seconds.

This one: 256K ctx · Mistral AI

MAI-Transcribe-1

A smart assistant you chat with — ask it to write, code, explain, summarize, or plan, and it answers in seconds.

This one: Microsoft

MAI-Voice-1

A smart assistant you chat with — ask it to write, code, explain, summarize, or plan, and it answers in seconds.

This one: Microsoft

MAI-Image-2

A smart assistant you chat with — ask it to write, code, explain, summarize, or plan, and it answers in seconds.

This one: Microsoft

MAI-Thinking-1

A smart assistant you chat with — ask it to write, code, explain, summarize, or plan, and it answers in seconds.

This one: 256K ctx · Microsoft

MAI-Code-1-Flash

A smart assistant you chat with — ask it to write, code, explain, summarize, or plan, and it answers in seconds.

This one: Microsoft

Command A+

A smart assistant you chat with — ask it to write, code, explain, summarize, or plan, and it answers in seconds.

This one: 128K ctx · Cohere

Nemotron 3 Super

A smart assistant you chat with — ask it to write, code, explain, summarize, or plan, and it answers in seconds.

This one: 1M ctx · NVIDIA

Nemotron 3 Ultra

A smart assistant you chat with — ask it to write, code, explain, summarize, or plan, and it answers in seconds.

This one: 262K ctx · NVIDIA

Granite 4.1

A smart assistant you chat with — ask it to write, code, explain, summarize, or plan, and it answers in seconds.

This one: 512K ctx · IBM

Stable Diffusion

Describe a picture in plain words and it draws it for you — no art skills needed.

This one: Open-weight, massive community ecosystem, self-hostable

DALL-E 3 / GPT Image

Describe a picture in plain words and it draws it for you — no art skills needed.

This one: Deep ChatGPT integration, being replaced by GPT Image 1/1.5

Midjourney

Type a few words and it paints strikingly artistic images.

This one: Industry-leading aesthetic quality, V7 architecture

Adobe Firefly

Describe a picture in plain words and it draws it for you — no art skills needed.

This one: Trained on licensed content (commercially safe), Creative Cloud integration

Ideogram

Describe a picture in plain words and it draws it for you — no art skills needed.

This one: Best-in-class text rendering in images, Magic Prompt

Flux 2 (Black Forest Labs)

Describe a picture in plain words and it draws it for you — no art skills needed.

This one: #1 Arena ELO (Pro v1.1), Klein generates in <0.5s, built-in typography

Leonardo AI AU

Describe a picture in plain words and it draws it for you — no art skills needed.

This one: Custom model training with 10-20 images, diverse style models

Jimeng / Seedream (ByteDance)

Describe a picture in plain words and it draws it for you — no art skills needed.

This one: 94% text rendering accuracy, 14 reference images, character consistency

Tongyi Wanxiang (Alibaba)

Describe a picture in plain words and it draws it for you — no art skills needed.

This one: First video model with CN/EN text effects, open-source Wan2.x video

Hunyuan Image/Video (Tencent)

Describe a picture in plain words and it draws it for you — no art skills needed.

This one: Largest open-source image model (80B), consumer GPU compatible, #1 LMArena

Recraft V3

Describe a picture in plain words and it draws it for you — no art skills needed.

This one: #1 on text-to-image ELO, vector/SVG native, brand consistency

Playground AI

Describe a picture in plain words and it draws it for you — no art skills needed.

This one: Mixed image editing canvas, combine multiple AI models

Imagen 3 (Google)

Describe a picture in plain words and it draws it for you — no art skills needed.

This one: Best-in-class photorealism, integrated with Gemini and Vertex AI

Midjourney V8

Describe a picture in plain words and it draws it for you — no art skills needed.

This one: Midjourney

Stable Diffusion 4

Describe a picture in plain words and it draws it for you — no art skills needed.

This one: Stability AI

GPT Image 2

Describe a picture in plain words and it draws it for you — no art skills needed.

This one: OpenAI

Midjourney V8.1

Describe a picture in plain words and it draws it for you — no art skills needed.

This one: Midjourney

HiDream-O1-Image

Describe a picture in plain words and it draws it for you — no art skills needed.

This one: HiDream-ai

Recraft V4.1

Describe a picture in plain words and it draws it for you — no art skills needed.

This one: Recraft

Ideogram 4.0

Describe a picture in plain words and it draws it for you — no art skills needed.

This one: Ideogram

Tripo Smart Mesh P1.0

Describe a picture in plain words and it draws it for you — no art skills needed.

This one: VAST AI

Tripo H3.1

Describe a picture in plain words and it draws it for you — no art skills needed.

This one: VAST AI

Hunyuan 3D World 2.0

Describe a picture in plain words and it draws it for you — no art skills needed.

This one: Tencent

Pixal3D

Describe a picture in plain words and it draws it for you — no art skills needed.

This one: TencentARC

Rodin Gen-2.5

Describe a picture in plain words and it draws it for you — no art skills needed.

This one: Deemos

Sora / Sora 2

Type a scene, or feed it a photo, and it generates moving video — like a tiny film studio.

This one: 25s video with synchronized dialogue/SFX/music from one prompt

Runway Gen-4

Type a scene, or feed it a photo, and it generates moving video — like a tiny film studio.

This one: Professional editing controls, in-video editing (Aleph), motion capture (Act-Two), 4K

Kling 3.0 (Kuaishou)

Type a scene, or feed it a photo, and it generates moving video — like a tiny film studio.

This one: Native 4K 60fps, up to 3-min video (longest), multi-language audio, 60M+ creators

Pika 2.5

Type a scene, or feed it a photo, and it generates moving video — like a tiny film studio.

This one: Physics-based interactions, Pikaswaps, Pika Selves (AI avatars), integrated SFX

VIDU Q3 (Shengshu)

Type a scene, or feed it a photo, and it generates moving video — like a tiny film studio.

This one: #1 VBench (above Sora), <10s generation, 55% cheaper than competitors

Seedance 2.0 (ByteDance)

Type a scene, or feed it a photo, and it generates moving video — like a tiny film studio.

This one: Unified audio-video joint generation, 2K cinema, 200+ digital avatars, native multi-shot consistency

HiggsField AI

Type a scene, or feed it a photo, and it generates moving video — like a tiny film studio.

This one: Multi-model video + talking avatars, face/body animation, social-first

Wan 2.2 (Alibaba)

Type a scene, or feed it a photo, and it generates moving video — like a tiny film studio.

This one: Apache 2.0, MoE flow-matching, best aesthetic quality (85.3), 14B params, self-hostable

LTX 2.3 (Lightricks)

Type a scene, or feed it a photo, and it generates moving video — like a tiny film studio.

This one: 19B params (14B video + 5B audio), 4K 50fps, native stereo audio, runs locally

Helios (ByteDance)

Type a scene, or feed it a photo, and it generates moving video — like a tiny film studio.

This one: 14B autoregressive diffusion, minute-long video (1440 frames), 19.5 FPS on single H100

Luma Dream Machine

Type a scene, or feed it a photo, and it generates moving video — like a tiny film studio.

This one: Fast generation, camera control, consistent characters, API access

Synthesia

Type a scene, or feed it a photo, and it generates moving video — like a tiny film studio.

This one: AI avatars in 140+ languages, enterprise training video leader

Hailuo AI (MiniMax)

Type a scene, or feed it a photo, and it generates moving video — like a tiny film studio.

This one: Free high-quality video gen, strong motion & character consistency

LTX-2.3

Type a scene, or feed it a photo, and it generates moving video — like a tiny film studio.

This one: Lightricks

Wan 2.7

Type a scene, or feed it a photo, and it generates moving video — like a tiny film studio.

This one: Alibaba

Gemini Omni Flash

Type a scene, or feed it a photo, and it generates moving video — like a tiny film studio.

This one: Google DeepMind

Runway Aleph 2.0

Type a scene, or feed it a photo, and it generates moving video — like a tiny film studio.

This one: Runway

NVIDIA Cosmos 3

Type a scene, or feed it a photo, and it generates moving video — like a tiny film studio.

This one: NVIDIA

Suno

Type a song idea and it produces a full track with vocals and instruments.

This one: V5 with built-in DAW (Suno Studio), stem editing, Warner Music partnership, licensed training

Udio

Give it a vibe, lyrics, or text and it makes music, a voice, or sound effects.

This one: 48kHz stereo up to 10 min, voice cloning, precise lyric control

ElevenLabs

Turns typed text into remarkably natural speech, and can clone a voice.

This one: Industry-leading voice quality, 70+ languages, professional voice cloning

Stable Audio 2.5

Give it a vibe, lyrics, or text and it makes music, a voice, or sound effects.

This one: Pro audio + SFX generation, partially open-source, Stability AI ecosystem

AIVA LU

Give it a vibe, lyrics, or text and it makes music, a voice, or sound effects.

This one: Classical composition, MIDI/MP3 export, full copyright transfer on paid plans

Soundraw

Give it a vibe, lyrics, or text and it makes music, a voice, or sound effects.

This one: Real-time customization (tempo, instruments, mood), royalty-free for all platforms

Boomy

Give it a vibe, lyrics, or text and it makes music, a voice, or sound effects.

This one: Create + release songs to Spotify/Apple Music, built-in distribution

PlayHT

Give it a vibe, lyrics, or text and it makes music, a voice, or sound effects.

This one: Most realistic TTS, voice cloning, emotion control, streaming API

Descript

Give it a vibe, lyrics, or text and it makes music, a voice, or sound effects.

This one: Edit audio/video by editing text transcript, AI voice cloning

Murf AI

Give it a vibe, lyrics, or text and it makes music, a voice, or sound effects.

This one: 200+ voices, 20+ languages, enterprise voice-over platform

Lyria 3 Pro

Give it a vibe, lyrics, or text and it makes music, a voice, or sound effects.

This one: Google DeepMind

Suno v5.5

Give it a vibe, lyrics, or text and it makes music, a voice, or sound effects.

This one: Suno

ACE-Step 1.5 XL

Give it a vibe, lyrics, or text and it makes music, a voice, or sound effects.

This one: ACE Studio

Stable Audio 3.0

Give it a vibe, lyrics, or text and it makes music, a voice, or sound effects.

This one: Stability AI

ElevenLabs Music v2

Give it a vibe, lyrics, or text and it makes music, a voice, or sound effects.

This one: ElevenLabs

GPT-Realtime-2

Give it a vibe, lyrics, or text and it makes music, a voice, or sound effects.

This one: OpenAI

GPT-Realtime-Translate

Give it a vibe, lyrics, or text and it makes music, a voice, or sound effects.

This one: OpenAI

OpenRouter

An all-in-one toolbox that bundles several AI tools behind a single login.

This one: 290+ models one API, no markup, automatic fallback routing

Poe (Quora)

An all-in-one toolbox that bundles several AI tools behind a single login.

This one: Multi-model chat in one app, custom bots, cheapest multi-AI access ($5/mo)

LM Studio

An all-in-one toolbox that bundles several AI tools behind a single login.

This one: Run AI 100% locally, OpenAI-compatible API, no data leaves your machine

SiliconFlow

An all-in-one toolbox that bundles several AI tools behind a single login.

This one: Ultra-cheap Chinese inference, 50+ models, faster than OpenAI in Asia

Together AI

An all-in-one toolbox that bundles several AI tools behind a single login.

This one: $100 free credits, open-source model hosting, serverless + dedicated GPUs

WaveSpeed AI

An all-in-one toolbox that bundles several AI tools behind a single login.

This one: Fastest Flux/SDXL inference, 2-4x faster than competitors, image+video gen API

Hugging Face Spaces

An all-in-one toolbox that bundles several AI tools behind a single login.

This one: 500K+ ML apps, largest model hub, community-driven, open-source ecosystem

Botpress

An all-in-one toolbox that bundles several AI tools behind a single login.

This one: Visual chatbot builder, GPT-powered agents, 500 free msgs/mo, deploy to any website

Elfsight

An all-in-one toolbox that bundles several AI tools behind a single login.

This one: 80+ embeddable widgets, AI chatbot widget, no-code setup, works on any website

Coze (ByteDance)

An all-in-one toolbox that bundles several AI tools behind a single login.

This one: Visual drag-and-drop agent builder, one-click deploy to Discord/WhatsApp/WeChat/Feishu, Vibe Coding

Nami AI (360 Group)

An all-in-one toolbox that bundles several AI tools behind a single login.

This one: 50+ LLMs freely switchable, #1 AI search in China, multi-agent swarm system, digital humans

Quark (Alibaba)

An all-in-one toolbox that bundles several AI tools behind a single login.

This one: 200M+ users, Qwen3-powered, AI browser + Super Box agent, 6TB free cloud, AI glasses

Notion AI

An all-in-one toolbox that bundles several AI tools behind a single login.

This one: AI integrated into workspace — summarize, write, Q&A over your docs

Jasper

An all-in-one toolbox that bundles several AI tools behind a single login.

This one: Marketing-focused AI — brand voice, campaign builder, multi-channel

Dify

An all-in-one toolbox that bundles several AI tools behind a single login.

This one: Visual AI app builder, RAG pipeline, agent workflows, open-source

The model runs on your own computer (via Ollama, LM Studio, or ComfyUI). Private — nothing leaves your device, no per-use fees, works offline. The catch: you're capped by your hardware, so big models need a strong GPU and run slower.

Best for privacy-sensitive work, tinkering, small/medium models, and avoiding usage bills.

The model runs on machines you reach via an API (Groq, Together, Replicate, fal) or a full cloud (AWS, Google Cloud, Azure). Runs the biggest models fast, scales to any traffic, nothing to install. The catch: data leaves your device and you pay per use.

Best for production apps, huge models, bursty traffic, and when you don't own a GPU.

Hugging Face

Model Hub

Open hub to host, share, and run AI models, datasets, and live demo apps (Spaces).

Civitai

Community Hub

The biggest open community hub for Stable Diffusion / Flux image-model resources — checkpoints, LoRAs, embeddings — plus on-site generation.

Replicate

Inference API

Run thousands of open + proprietary models (image, video, audio, language) via one API — no GPU to manage.

fal.ai

Inference API

Serverless inference for 1,000+ optimized generative models (image, video, audio, 3D) through one fast API.

Modal

Serverless GPU

Run Python (especially AI/ML) on auto-scaling GPUs and CPUs, billed per-second, no infra to manage.

ComfyUI

Node Workflow

Open-source node-based tool for building local image/video generation pipelines by wiring nodes — no code.

Ollama

Local Runtime

Download and run open LLMs locally with one command; exposes an OpenAI-compatible API at localhost.

Together AI

Inference API

Run and fine-tune open-source models (Llama, DeepSeek, Qwen, GPT-OSS, 200+) behind a fast OpenAI-compatible API.

Groq

Inference API

Run open-source LLMs at very high speed and low cost on custom LPU hardware (not GPUs).

RunPod

GPU Cloud

Rent NVIDIA GPUs by the second — full container Pods you control, or auto-scaling Serverless inference endpoints.

Amazon Web Services (AWS)

Cloud & Compute

The biggest cloud — rent servers, GPUs, storage, and managed AI (Bedrock, SageMaker) at any scale.

Google Cloud (GCP)

Cloud & Compute

Google's cloud with Vertex AI, NVIDIA GPUs and Google's own TPUs, plus direct access to Gemini.

Microsoft Azure

Cloud & Compute

Microsoft's enterprise cloud, plus Azure AI Foundry hosting frontier models (incl. OpenAI) with compliance.

Lambda

Cloud & Compute

A GPU cloud built specifically for AI — on-demand and reserved NVIDIA GPUs at developer-friendly prices.

CoreWeave

Cloud & Compute

A specialized GPU cloud built for large-scale AI training and high-throughput inference.

Gemini Live APIGoogle

The only cloud API doing true continuous webcam/screen video streaming with voice — interruption and turn-taking built in.

• ~1fps video sampling · sub-second voice replies

OpenAI gpt-realtimeOpenAI

Production voice agents over WebRTC that can glance at the camera — frames are snapshotted as billed images, not continuous video.

• ~1fps snapshots · sub-second audio

Qwen3-OmniAlibaba

Open-source Gemini-Live equivalent: streams audio+video in and speaks back in real time. Self-hostable for privacy.

• 507ms audio+video e2e (paper) · Apache-2.0

MiniCPM-o 4.5OpenBMB

Full-duplex multimodal live streaming on-device — continuous video+audio in, speech out, demoed on an iPad.

• up to 10fps video (claim) · ~5-6GB int4 (est)

SmolVLM2 realtimeHF

Tiny Apache VLM whose one-command llama.cpp demo made local realtime webcam captioning a laptop thing.

• runs realtime on MacBook · <1GB RAM (est)

Moondream 3US

Always-on small vision model with detection, pointing and gaze skills — Moondream 2 runs on a Raspberry Pi.

• 9B MoE / 2B active · M2 ~1.2GB int4

MediaPipe TasksGoogle

The reflex layer under nearly every gesture, avatar and posture app: 478-pt face, 21-pt hands, 33-pt pose at 30-60fps.

• 12-17ms per frame (doc'd) · Apache-2.0

YOLO26Ultralytics

Current realtime object-detection flagship for what's-in-frame events feeding VLMs or automations.

• 1.7-11.8ms on T4 (doc'd) · AGPL — mind the license

KalidokitOpen

Solver turning MediaPipe landmarks into VRM/Live2D avatar rigs — the core of browser VTuber apps.

• MIT · runs at tracker speed in-browser

Project GamefaceGoogle

Open-source hands-free control: head movement + facial gestures become your mouse, on desktop and Android.

• webcam-only · Apache-2.0

dots.ocrOpen

Compact open OCR model that tops document-parsing leaderboards; great layout + tables.

PaddleOCR-VLBaidu

Vision-language OCR handling 100+ languages, handwriting, formulas and tables.

olmOCR 2AI2

Fully open OCR pipeline for turning PDFs and scans into clean, structured text.

Mistral OCRFrance

Fast document-understanding API that preserves layout, images and math markdown.

DeepSeek-OCRChina

Context-compression OCR that packs long documents into few vision tokens.

Meshy 6US

Text/image-to-3D with clean topology and PBR textures; the prosumer 3D-asset default.

Tripo v3US

Fast, detailed 3D generation with rigging and part segmentation for games.

Rodin Gen-2Deemos

High-fidelity 3D characters and props with sharp geometry and materials.

TRELLIS.2Microsoft

Open image-to-3D model producing detailed meshes and radiance fields.

Hunyuan3D-3Tencent

Open, high-resolution 3D generation with strong texture synthesis.

MedGemmaGoogle

Open medical vision-language models for imaging and clinical text (research use).

• Not a medical device — not for diagnosis.

Med-GeminiGoogle

State-of-the-art on medical QA (USMLE-style); multimodal clinical reasoning.

• Research model, not clinical advice.

OpenEvidenceUS

Point-of-care assistant for clinicians, grounded in peer-reviewed literature.

• For licensed clinicians.

Google AMIEGoogle

Research diagnostic-dialogue agent studied for history-taking and reasoning.

• Research only.

Gemini EmbeddingGoogle

Top-ranked general-purpose text embeddings for search and RAG.

Qwen3-EmbeddingAlibaba

Strong open multilingual embeddings and rerankers for retrieval.

Voyage-3US

High-accuracy embeddings tuned for code and domain retrieval.

NV-EmbedNVIDIA

Research-grade embeddings topping the MTEB retrieval leaderboard.

Genie 3Google

Generative world model that turns a prompt into an explorable, interactive environment.

NVIDIA GR00TNVIDIA

Foundation model for humanoid robots — perception, reasoning and motor control.

1X World Model1X

Learned simulator for training and evaluating home-robot behavior.

Claude CodeAnthropic

Terminal-native agent; plans, edits, runs tests and subagents. Backend: Claude Fable 5 / Opus.

• Skill: parallel subagents + skills/hooks.

CursorUS

AI-first editor whose Composer 2.5 agent writes whole features while you watch.

• Skill: in-house fast agent model.

Kimi (K3 + Kimi Work)Moonshot

K3 (Jul 2026) with a 300-agent swarm and Kimi Work desktop agent; FrontierSWE ~81.

• Skill: massive parallel agent swarm.

ChatGPT AgentOpenAI

Computer-use agent that browses, runs code in a sandbox and completes multi-step tasks.

• Skill: unified browser + terminal tool-use.

Gemini AgentGoogle

Agentic mode in Gemini; Project Mariner was folded in after shutting down May 2026.

• Skill: deep Google/Workspace integration.

WindsurfUS

Agentic IDE with its own fast SWE-1.5 model and flow-style multi-file edits.

• Skill: in-house SWE-1.5 model.

DevinCognition

Autonomous software engineer that takes a ticket and ships a PR end-to-end.

• Skill: long-horizon autonomous tasks.

GitHub CopilotGitHub

Agent mode plans and edits across a repo; strongest as inline autocomplete.

• Skill: repo-wide agent + autocomplete.

ClineOpen

Open-source VS Code agent; bring your own model key, full tool-use transparency.

• Skill: open, model-agnostic.

AmpSourcegraph

Agentic coding tool built for large codebases with deep code search.

• Skill: codebase-scale context.

Roo CodeOpen

Popular open agent fork — discontinued May 15, 2026. Community forks continue.

• Discontinued.

Plaud NotePinUS

Wearable recorder that transcribes and summarizes every meeting and idea.

• $169 + plan

Limitless PendantUS

Always-listening pendant that builds a searchable memory of your day.

• $99 + plan

Bee AIUS

Wristband life-recorder that turns conversations into to-dos and notes (Amazon-acquired).

• $50

CompassUS

Pin-style AI companion that listens and coaches through your day.

• wearable

Ray-Ban Meta (Gen 2)Meta

The best-selling AI glasses — camera, open-ear audio and live Meta AI, now with a display.

• from $379

Oakley Meta VanguardMeta

Sport-focused Meta AI glasses with a wide camera for cycling and running.

• from $499

Even Realities G1US

Minimalist glasses with a discreet display for navigation, notes and translation.

• ~$599

Xreal One ProChina

AR display glasses that put a giant virtual screen in front of you.

• ~$649

Halliday GlassesUS

Proactive AI glasses with an invisible near-eye display and ring control.

• ~$499

Timekettle EarbudsChina

Real-time translation earbuds that interpret face-to-face across 40+ languages.

• translation

iFlytek AI EarbudsChina

Recording + translation earbuds tied to a strong speech-AI engine.

• translation

Apple AirPods (Live Translation)Apple

Mainstream earbuds now doing on-device live translation and health sensing.

• from $179

Oura Ring 4Finland

Smart ring with an AI advisor that reads sleep, recovery and readiness daily.

• $349 + plan

Whoop 5.0US

Screen-free band whose AI coach turns strain and recovery into daily guidance.

• membership

Ultrahuman Ring AirIndia

No-subscription smart ring with metabolic and sleep AI insights.

• $349

Eight Sleep Pod 5US

AI mattress cover that auto-adjusts temperature and tracks sleep all night.

• premium

Copilot+ PCsMicrosoft

Windows laptops with an NPU for on-device AI — Recall, live captions, local models.

• from $799

NVIDIA DGX SparkNVIDIA

Desktop AI supercomputer that runs and fine-tunes large models locally.

• ~$3,999

Jetson ThorNVIDIA

Robotics dev kit powering on-device humanoid and edge-AI brains.

• dev kit

Samsung Galaxy AI phonesSamsung

Flagship phones with live translation, generative photo edit and an on-device assistant.

• flagship

Alexa+Amazon

LLM-rebuilt Alexa that holds real conversations and runs multi-step home routines.

• $19.99/mo or Prime

Google Home (Gemini)Google

Gemini-powered home assistant across Nest speakers, displays and cameras.

• Nest devices

Aqara AI HubChina

Local-AI smart-home hub with camera reasoning and natural-language automations.

• hub

Figure 03US

Humanoid robot for the home, driven by a Helix vision-language-action model.

• humanoid

1X Neo1X

Soft, home-safe humanoid designed to help with chores around the house.

• ~$20k / sub

Unitree G1 / R1China

Affordable, agile humanoid robots popular with developers and researchers.

• from ~$5,900

Roborock Saros Z70China

Robot vacuum with a folding arm that picks up socks and clutter as it cleans.

• premium

RopetUS

Fluffy AI desk pet with expressive eyes that reacts to touch and mood.

• companion

Casio MoflinCasio

AI comfort pet that develops its own personality the more you hold it.

• ~$430

MirokiFrance

Expressive companion robot from Enchanted Tools for homes and hospitality.

• companion

Tamagotchi (AI era)Bandai

The classic digital pet, reimagined with AI-driven personality and chat.

• toy

Hume AI (Octave 2 / EVI)US

Emotion-intelligent voice: Octave 2 TTS and EVI empathic speech-to-speech that reads and renders 48+ emotions.

• ✓ Creator $14/mo · free 10k chars + 5 EVI min/mo

Fish Audio (S2.1 Pro)US

Expressive realtime TTS with 15-second voice cloning and emotion tags in 30+ languages.

• ✓ Plus $15/mo · API $15/M bytes · free flagship tier

Cartesia (Sonic-3.5)US

Ultra-low-latency streaming TTS (~40ms) on a state-space architecture, plus Ink STT and the Line agent platform.

• ✓ Pro $5/mo · free 20k credits/mo

ElevenLabs (v3)US

The category default: expressive TTS, cloning, dubbing, music and a full agents platform.

• from $5/mo · free monthly credits

MiniMax Audio (Speech 2.6)CN

Arena-topping expressive multilingual TTS API with instant voice cloning.

• per-character API

Inworld TTSUS

Frontier-quality TTS priced aggressively for scale, from the AI-character engine company.

• low per-character pricing

Resemble AI (Chatterbox)CA

Enterprise cloning + speech-to-speech with watermarking and deepfake detection; open-source Chatterbox (MIT).

• Chatterbox open weights (MIT)

KokoroOpen

82M-parameter open TTS with near-commercial quality that runs on a plain CPU.

• free weights (Apache-2.0)

Neuphonic (NeuTTS Air)UK

Fast low-cost TTS API plus an open sub-1B on-device voice-cloning model.

• NeuTTS Air open weights

Rime (Arcana)US

Realtime TTS tuned for high-volume phone calls with demographically diverse casual voices.

• usage-based API

LMNTUS

Low-latency streaming TTS with instant cloning, aimed at realtime apps and games.

• usage-based API

PlayHT / PlayAIUS

Voice cloning and TTS platform — core team acquired by Meta (Jul 2025); future uncertain.

• acquired by Meta — check status

Deepgram (Nova-3 / Flux)US

Realtime STT plus Flux, a turn-aware conversational model built for voice agents, and Aura-2 TTS.

• ✓ Nova-3 from $0.0048/min · $200 signup credit

AssemblyAI (Universal-3.5)US

STT plus audio intelligence: summarization, PII redaction and sentiment on top of transcripts.

• ✓ $0.21/hr async · $50 signup credits

Speechmatics (Ursa)UK

Enterprise multilingual STT with realtime diarization and best-in-class accent robustness; on-prem options.

• PAYG + free monthly hours

Gladia (Solaria)FR

EU-hosted STT with realtime streaming, code-switching and 100 languages, built for call platforms.

• usage-based API

VapiUS

Developer platform to build and scale phone/web voice agents on any model stack.

• ✓ $0.05/min platform fee + components at cost

Retell AIUS

Voice-agent platform for call ops: build, simulate and deploy phone agents with bundled telephony.

• ✓ $0.07-0.31/min all-in · $10 credit

Bland AIUS

Enterprise AI phone-calling on its own end-to-end self-hosted model stack for latency and compliance.

• per-minute PAYG

LiveKit AgentsUS

Open-source WebRTC infrastructure + agents framework; powers ChatGPT Voice.

• OSS free · managed cloud plans

Pipecat (Daily)US

Vendor-neutral open-source Python framework for realtime voice and multimodal agents.

• OSS free (BSD-2) · hosted cloud

HeyGen (Avatar IV/V)US

AI avatar video studio and API: digital twins, Video Agent, interactive avatars in 175+ languages.

• ✓ Creator $29/mo · free 3 videos/mo

Tavus (Phoenix-4)US

Conversational video API — realtime face-to-face AI humans with perception and turn-taking models.

• ✓ Starter $59/mo (100 min) · free 25 min/mo

SynthesiaUK

Enterprise AI video standard: 230+ stock avatars and 140+ languages for training content.

• ✓ Starter $29/mo · free 10 min/mo

Hedra (Character-3)US

Omnimodal character animation: make any image talk or sing expressively — stylized characters included.

• subscription + free credits

D-IDIL

Photo-to-talking-video pioneer with lightweight embeddable realtime web agents.

• subscription tiers

Sync (lipsync-2)US

Zero-shot lipsync API that re-times any video's mouth to new audio in any language — from the Wav2Lip authors.

• per-minute API

Mem0US

Memory-as-a-service: agents get persistent, self-improving user memory across sessions.

• OSS core + hosted tiers

LettaUS

Stateful agents from the MemGPT team — long-lived assistants that manage their own memory.

• OSS (Apache-2.0) + cloud

Browserbase (Stagehand)US

Headless-browser fleets for AI agents, with the Stagehand act/extract automation framework.

• free dev browser · paid fleets

Browser UseOpen

The 60k-star open-source library letting LLM agents drive real browsers; hosted cloud available.

• OSS free (MIT)

E2BUS

Firecracker sandboxes agents spin up in ~150ms to safely run untrusted code.

• usage-based + hobby tier

ExaUS

Neural web-search API built for LLMs — semantic search over its own embeddings-native index.

• PAYG + starter credit

TavilyIL

Search API optimized for RAG and agents; the default search tool in LangChain examples.

• free monthly credits

FirecrawlUS

Scraping API that turns any website into clean, LLM-ready markdown or structured JSON.

• free starter credits

NVIDIA RTX 5090NVIDIA

Flagship consumer GPU (32GB GDDR7). Runs 32B models fully in VRAM, 70B with CPU offload. Street price ~2x MSRP in the memory shortage.

• $3,999 street (MSRP $1,999) · verified 2026-07-24, Tom's Hardware

NVIDIA RTX 5080NVIDIA

16GB GDDR7 upper-tier card — the value local-AI pick of the 50 series. Comfortable with 14B models, 32B at aggressive quantization.

• $1,289 street (MSRP $999) · verified 2026-07-24, Tom's Hardware

NVIDIA RTX 4090NVIDIA

Prior-gen 24GB workhorse; new units now cost MORE than launch and used eBay averages ~$2,270 — the used-bargain era is over.

• $3,489 new / ~$2,270 used · verified 2026-07-24, Tom's Hardware + BestValueGPU

AMD Radeon RX 9070 XTAMD

AMD's consumer flagship (RDNA 4, 16GB). Runs 14B models via ROCm/Vulkan in llama.cpp or LM Studio.

• $689 street (MSRP $599) · verified 2026-07-24, Tom's Hardware

AMD Radeon AI PRO R9700AMD

32GB workstation card aimed squarely at local AI — the most VRAM per dollar of any new card. 32B at high quality, 70B across two cards.

• $1,299 MSRP · verified 2026-05-31, AMD / GPU Poet

NVIDIA DGX SparkNVIDIA

GB10 desktop AI supercomputer with 128GB unified memory — runs 70B-120B models locally, ~200B across two linked units. MSRP raised $700 in Feb 2026.

• $4,699 MSRP (was $3,999) · verified 2026-07-18, NVIDIA

Mac Studio (M4 Max / M3 Ultra)Apple

The local-LLM darling: from 36GB up to 512GB unified memory (M3 Ultra), which runs 670B-class models quantized via MLX.

• from $2,499 · verified 2026-07-26, Apple Store

Mac mini (M4)Apple

Cheapest capable Mac for local models — 8B on the 16GB base, 14B-32B on M4 Pro configs. Base rose $200 in Apple's Jan-2026 reprice.

• from $799 (was $599) · verified 2026-07-26, Apple Store

Jetson AGX Thor Dev KitNVIDIA

Blackwell robotics/edge dev kit: 2,070 FP4 TFLOPS and 128GB in a 40-130W box — 70B+ at the edge, built for robot brains.

• $3,499 MSRP · verified 2026-07-18, NVIDIA resellers

Framework Desktop (128GB)Framework

Strix Halo mini-PC with 128GB unified memory — the value alternative to DGX Spark; runs 70B and gpt-oss-120b via Ollama/llama.cpp.

• $1,999 base · verified 2026-07-18, Framework

Jetson Orin Nano SuperNVIDIA

The $249 entry ticket to NVIDIA edge AI: 67 TOPS, 8GB, full CUDA stack — runs 3-8B quantized LLMs and vision pipelines.

• $249 · verified 2026-07-18, NVIDIA (in stock)

Raspberry Pi AI HAT+ (26 TOPS)RPi

Official Hailo-8 accelerator for Pi 5 — the default hobbyist edge-vision kit for YOLO, pose and segmentation (not LLMs).

• $110 ($70 for 13-TOPS) · verified 2026-07-18, Raspberry Pi