


Playable browser games from a single prompt.

Claude Fable 5.2

GPT-6.2 Astra

GPT-6 Astra
(max)
Claude Opus 5.5

GPT-6.2 Sol
Claude Sonnet 5.5

Inkling 2 Small

Gemini 4 Carbon

Qwen4-Max

Muse Wonder
(max)

Gemini 4 Neon

Kimi K3.1 Pro

Llama 4 Scout
Evaluates model ability to generate playable browser games as single-file HTML from real-world user prompts. Real-world users interact with the outputs and select their preferred result.


















































































Updated Nov 21, 2026, 6:33 AM UTC. Elo Rating: A skill-based rating system starting at 1200. Ratings are calculated using the standard Bradley-Terry model. See more in our Methodology page. Margin of Error: Shows ±standard error from the Bradley-Terry model for Elo rating confidence intervals.