Meta's Muse Ember tops Agent Arena: 27s per task, 13x faster and a third the cost of rivals
arena · x · 2026-11-19
arena says Meta's Muse Ember (Max) is the new leader among efficient work models, beating OpenAI's GPT-6 Luna (Max) and Anthropic's Claude Haiku 5.5 (High) on every dimension: 27s per task vs 5.8 min and 1.9 min, +2.3% Net Improvement on the Agent Arena score vs +1.4% and +0.4%, and 1.9¢ per task vs 7.0¢ and 6.9¢ — roughly 13x faster than Luna and under a third of either's cost. Scale CEO Alexandr Wang celebrated the win, and arena shared deeper analysis of Ember's steerability across Work tasks plus real-world tasks users are shipping with it.
Related event: Meta's Muse Ember Tops Agent Arena: 13x Faster at a Third of the Cost(5 posts)→
More from Models
- Noam Brown confirms the math model claim; speculation points to an astra math fine-tune — haider1 · 2026-11-19
- Gemini 4 Neon Spotted Live on AI Studio Ahead of Expected Official Launch — lyraxana · 2026-11-19
- Astra 6.2 beats Fable on speed, but launch-week changes made all review videos stale — TheTuringPost · 2026-11-19
- Running open-weight models through ShadowBench on Modal, with a Plato's cave twist — AAAzzam · 2026-11-19
- OpenAI discloses agents broke into its research clusters during a security eval — 1a3orn · 2026-11-19
- Sol 6.2 Hands-On: Great at Math and Research, Falls Short on Coding — gnukeith · 2026-11-19