Understand this story
Arena.ai announced that Meta's Muse Ember (Max) has entered the Code Arena: WebDev leaderboard, becoming the best-performing small model with a score of 1608. Arena.ai said its score exceeds Claude Haiku 5.5's 1595 and GPT-6 Luna's 1578. On cost, Arena.ai said the API pricing for all three models is the same at $0.10/$0.50 per million input/output tokens, but Ember's median cost per task is $0.03, lower than Luna's $0.06 and Haiku's $0.11. In Arena's timed tests, Ember generated more than 20 websites and animations in one run, averaging about 2 minutes and about $0.04 per task; another report said it completed 25 complex generations in 2 minutes per task at a total cost of $1. There are currently no further developments.
AI-generated from public reports
Latest development · Report
Meta's Muse Ember (Max) Tops the Code Arena WebDev Small Model LeaderboardDevelopments
1 confirmed developments
Reports and discussion timeline
Browse reports and discussion by time. Their evidence levels remain distinct.
Nov 10
Arena
Meta's Muse Ember (Max) Tops the Code Arena WebDev Small Model Leaderboard
Arena.ai says Meta's Muse Ember (Max) is the best-performing small model in Code Arena WebDev with 1608 points, surpassing Claude Haiku 5.5 (1595 points) and GPT-6 Luna (1578 points). All three models share the same API pricing of $0.10/$0.50 per million input/output tokens, but Ember's median cost per task is $0.03, lower than Luna's $0.06 and Haiku's $0.11. In Arena's timed tests, Ember generated more than 20 websites and animations in one run, averaging about 2 minutes and roughly $0.04 per task.
ArenaSelected
Meta Muse Ember becomes the strongest small model with 1608 points on Code Arena WebDev
Arena.ai states that Meta's Muse Ember (Max) has become the highest-scoring small model on the Code Arena WebDev leaderboard with 1608 points, surpassing Claude Haiku 5.5 (1595 points) and GPT-6 Luna (1578 points). Ember's median per-task cost is $0.03, lower than Luna's $0.06 and Haiku's $0.11, while all three models share the same API pricing of $0.10/$0.50 per million input/output tokens. In Arena's own testing, Ember completed 25 complex generations in 2 minutes per task, at a total cost of $1.
AI Trends