Meta launches Muse Ember, its smallest model built for per-dollar agent intelligence

On November 10, Meta officially released Muse Ember, its smallest and most efficient model to date, positioned as an efficiency-focused model for round-the-clock repetitive agent work. Its core selling point is "intelligence per dollar"—Meta claims that at any inference tier, its intelligence output per unit cost exceeds that of Claude Haiku 5.5, Luna, and various flash models. Scale AI founder Alexandr Wang also announced the release the same day, stating that Muse Ember costs only one-third as much as comparable models.

Confirmed

• The release was confirmed via Meta's official blog and the @AIatMeta account, with Zuckerberg also taking part in the announcement
• Model positioning: everyday agent tasks such as tool calling, coding subtasks, routing, information extraction, and long-document parsing; it can carry workflows across applications, call tools in sequence, run tests, and make targeted code fixes
• Features a 1 million token context
• Official figures: outperforms Claude Haiku 5.5 on key agent and coding benchmarks, at a cost of $0.116 per task—roughly one-third of Haiku's cost
• Available through Meta's Model open channel (per reposted posts)

Not yet confirmed

• The comparison methodology and specific benchmark details behind the claim of "exceeding Haiku, Luna, and all flash-class models": so far only official statements and secondhand accounts exist, with no independent evaluations
• The exact division of labor with Scale AI: Alexandr Wang participated in the announcement, but the post does not specify Scale AI's role in the model

Why it matters

• The cost bottleneck for running agents at scale lies in inference overhead from high-frequency small tasks. Muse Ember targets this scenario at roughly one-third the cost—if the numbers hold, it will directly threaten the market share of Haiku and flash-class models
• Combining a 1 million token context with tool-calling capability lets even a small model handle complex pipeline work like long-document parsing, marking a further split toward the "small model + specialization" route

2026-11-10 ~ 2026-11-10 · 15 related posts

Full story(2 episodes)→

• Episode 1: Meta launches Muse Ember, its smallest model built for per-dollar agent intelligence(2026-11-10, 15 posts)
• Episode 2: Artificial Analysis Reviews Meta's Muse Ember: Matches Haiku 5.5 at One-Third the Cost(2026-11-10, 17 posts)

• Zuckerberg releases Muse Ember, Meta's smallest model touting higher intelligence per dollar — AIatMeta · 2026-11-10
• [source] Scale AI launches Muse Ember, a small model claiming Haiku-level intelligence at 1/3 the cost — alexandr_wang · 2026-11-10
• [source] Meta launches Muse Ember, its smallest model with 1M context built for agent work — AIatMeta · 2026-11-10
• Meta's Muse Ember beats Claude Haiku 5.5 on agent benchmarks at one-third the cost — kimmonismus · 2026-11-10
• Meta launches Muse Ember: $0.116 per task, sits on the efficiency frontier — kimmonismus · 2026-11-10
• Meta launches Muse Ember, its smallest model, claiming 3x cheaper than Haiku-class rivals — stablequan · 2026-11-10
• [source] Muse Ember debuts: near Spark 1.3 intelligence at 17x lower price, first Wonder Full-trained model — alexandr_wang · 2026-11-10
• Meta researcher touts Muse Ember, a small model for agents that beats Haiku on cost-efficiency — shengjia_zhao · 2026-11-10
• Reddit discussion: Meta's Muse Ember, its smallest and most efficient model — Recoil42 · 2026-11-10
• Meta launches Muse Ember: Haiku-level intelligence at 1/3-1/2 the price — Bitter-College8786 · 2026-11-10
• Alexandr Wang launches Muse Ember: near-max intelligence at 17x cheaper price — kimmonismus · 2026-11-10
• Meta's Ember beats Haiku 5.5 and Luna at $0.10/1M input pricing — YashasGunderia · 2026-11-10

3 near-duplicate retellings: alexandr_wang · WenhuChen · AIatMeta