Try every frontier AI model

39 models across chat, image, video, and music — one account, pay only for what you use.

Chat & reasoning

Claude Opus 5 logo

Claude Opus 5

Chat & reasoning

Anthropic's everyday flagship for complex agentic coding and enterprise work. Ranks first overall on the AA Intelligence Index v4.1 at 61 with max effort.

Try Claude Opus 5
Claude Fable 5 logo

Claude Fable 5

Chat & reasoning

Anthropic's highest-capability model for the hardest reasoning and long-horizon agentic work. Scores 60 on the AA Intelligence Index v4.1, just behind Opus 5 at max effort.

Try Claude Fable 5
GPT-5.6 Sol logo

GPT-5.6 Sol

Chat & reasoning

OpenAI's new flagship. AA Intelligence Index v4.1 of 58.9 and a new state of the art on the Coding Agent Index (80) — the top performance-per-dollar frontier model, with an `ultra` multi-agent mode for the hardest work.

Try GPT-5.6 Sol
Claude Opus 4.8 logo

Claude Opus 4.8

Chat & reasoning

Frontier all-rounder. Leads SWE-bench Pro (69%) at AA Intelligence Index v4.1 of 56 — the value pick just below Fable 5.

Try Claude Opus 4.8
GPT-5.6 Terra logo

GPT-5.6 Terra

Chat & reasoning

Balanced GPT-5.6 tier — matches GPT-5.5 quality (AA Intelligence Index 55) at roughly half the cost, tuned for everyday professional work.

Try GPT-5.6 Terra
GPT-5.5 logo

GPT-5.5

Chat & reasoning

OpenAI's previous flagship. AA Intelligence Index v4.1 of 54.8 — a strong all-round frontier model, now succeeded by the GPT-5.6 family.

Try GPT-5.5
Grok 4.5 logo

Grok 4.5

Chat & reasoning

xAI's smartest model, trained alongside Cursor for coding and agentic work. 500K context, always-on reasoning, live X knowledge, at a competitive price.

Try Grok 4.5
Claude Sonnet 5 logo

Claude Sonnet 5

Chat & reasoning

Anthropic's current fast workhorse — near-Opus quality at much lower latency and cost.

Try Claude Sonnet 5
GPT-5.6 Luna logo

GPT-5.6 Luna

Chat & reasoning

Fastest, most affordable GPT-5.6 tier — outperforms Claude Opus 4.8 on the Coding Agent Index at roughly a quarter of the cost.

Try GPT-5.6 Luna
GLM-5.2 logo

GLM-5.2

Chat & reasoning

Zhipu's GLM-5.2 — the leading open-weights model on the AA Intelligence Index.

Try GLM-5.2
Muse Spark 1.1 logo

Muse Spark 1.1

Chat & reasoning

Meta Superintelligence Labs' multimodal reasoning model, built for agentic work. 1M-token context, image + video understanding, search grounding, and strong coding — served on the new OpenAI-compatible Meta Model API at value pricing.

Try Muse Spark 1.1
Gemini 3.5 Flash logo

Gemini 3.5 Flash

Chat & reasoning

Fast, cheap frontier multimodal — 1M context, strong for high-volume production.

Try Gemini 3.5 Flash
Gemini 3.6 Flash logo

Gemini 3.6 Flash

Chat & reasoning

Google's fast agentic workhorse — 1M context, stronger coding and multimodal quality with more efficient output.

Try Gemini 3.6 Flash
Gemini 3.1 Pro logo

Gemini 3.1 Pro

Chat & reasoning

Google DeepMind frontier. 2M context, fully multimodal, and an AA Intelligence Index v4.1 score of 46.

Try Gemini 3.1 Pro
MiniMax M3 logo

MiniMax M3

Chat & reasoning

MiniMax M3 — Apache-licensed, matches Gemini 3.1 Pro on coding at open-weights pricing.

Try MiniMax M3
DeepSeek V4 Pro logo

DeepSeek V4 Pro

Chat & reasoning

Flagship open-source MoE (1.6T) — frontier reasoning + coding with a 1M-token context, at open-weights pricing.

Try DeepSeek V4 Pro
Kimi K2.6 logo

Kimi K2.6

Chat & reasoning

Moonshot's K2.6 — open-weights, 80%+ SWE-bench Verified, top-tier agentic reasoning.

Try Kimi K2.6
Qwen 3.7 Plus logo

Qwen 3.7 Plus

Chat & reasoning

Alibaba's flagship closed model — strong reasoning + function calling, served exclusively on Fireworks outside Alibaba's own cloud.

Try Qwen 3.7 Plus
Grok 4.3 logo

Grok 4.3

Chat & reasoning

xAI flagship. Configurable reasoning, 1M context, and optional live web knowledge at low frontier pricing.

Try Grok 4.3
GPT-5 logo

GPT-5

Chat & reasoning

OpenAI workhorse — strong general purpose at lower cost than 5.5.

Try GPT-5

Not sure which to pick? Read our guide to the best AI model for coding or compare models head-to-head on TryAI.