Radar Ultra

Model

X @arena

Arena.ai says Muse Ember (Max) is the top-performing small model in Code Arena: WebDev

Arena.ai said Muse Ember (Max) scored 1,608 in Code Arena: WebDev, making it the top-performing small model; Claude Haiku 5.5 (High) scored 1,595 at No. 35, and GPT-6 Luna (Max) scored 1,578 at No. 42. Arena.ai reported an average of about two minutes and $0.04 per task across timed generations; its reported median task cost was $0.03.

Arena.ai said all three models had the same API prices—$0.10 per million input tokens and $0.50 per million output tokens—while Ember’s median task cost was $0.03, versus $0.06 for GPT-6 Luna (Max) and $0.11 for Claude Haiku 5.5 (High). Arena.ai also said its own test completed 25 complex generations at about two minutes per task for $1 total, and showed Ember generating 20+ websites and animations in one go.

Written by AI from the reports · updated 5 days ago

Sources

Radar Ultra
An unofficial tool, not affiliated with OpenAI. Compiled from public posts; the original posts are authoritative.