The world's most-used AI assistant
🧠 9.5
💻 9.2
🖉 9.2
🤖
Claude 4 (Sonnet/Opus)
Anthropic
Best for Writing
Helpful, harmless, and honest AI
🧠 9.4
💻 9.0
🖉 9.6
#1 on SWE-bench Pro (80.4%) — Anthropic's most capable model
🧠 —
💻 —
🖉 —
🤖
Claude Sonnet 5
Anthropic
Anthropic's mid-tier workhorse — $2/$10/M intro pricing through August 31, 2026
🧠 —
💻 —
🖉 —
Open-source model rivaling GPT-4
🧠 9.0
💻 9.2
🖉 8.5
$0.44/$0.87/M discounted — most cost-efficient frontier coding model. API migration deadline July 24.
🧠 —
💻 —
🖉 —
🤖
Gemini 3.6 Flash
Google DeepMind
$1.50/$7.50/M — 17% fewer output tokens, 304 tok/s, Computer Use built in
🧠 —
💻 —
🖉 —
Google's most capable multimodal model
🧠 9.3
💻 8.8
🖉 8.8
🤖
Gemma 4 (31B / 26B MoE)
Google DeepMind
Best Free Local Model
Byte for byte, the most capable open models
🧠 9.0
💻 9.0
🖉 8.8
OpenAI's frontier model — AA Index #1, $5/$30/M
🧠 —
💻 —
🖉 —
Witty, real-time AI by Elon Musk's xAI
🧠 8.8
💻 8.5
🖉 8.5
xAI's frontier model — AA Index #2, $2/$6/M, 64.7% SWE-bench Pro
🧠 —
💻 —
🖉 —
2.8T params, $3/$15/M, AA Index #4 — Design Arena #1 and SWE Marathon #1
🧠 —
💻 —
🖉 —
Meta's open-source frontier model
🧠 8.5
💻 8.5
🖉 8.2
Free AI assistant built into WhatsApp, Instagram, and Facebook
🧠 —
💻 —
🖉 —
🤖
Microsoft Copilot
Microsoft
Your everyday AI companion
🧠 8.5
💻 8.0
🖉 8.5
🤖
Character.AI
Character.AI
Roleplay and chat with AI characters
🧠 7.5
💻 4.0
🖉 9.0
Built for enterprise RAG and agents
🧠 8.0
💻 7.5
🖉 8.0
UAE's open-source frontier model
🧠 8.0
💻 7.8
🖉 7.8
Google's open-weights research model
🧠 7.8
💻 7.5
🖉 7.8
Frontier model with strong reasoning
🧠 8.5
💻 8.8
🖉 8.2
Compact but surprisingly capable
🧠 7.5
💻 7.8
🖉 7.5
Your personal AI built for empathy
🧠 7.5
💻 4.5
🖉 8.5
Chat with 50+ AI models in one app
🧠 7.5
💻 7.0
🖉 8.2
Alibaba's multilingual open model
🧠 8.5
💻 8.5
🖉 8.2
Kai-Fu Lee's bilingual frontier model
🧠 8.0
💻 7.8
🖉 8.0