MODEL

Muse Ember

modeltopic-notemeta

Overview

Muse Ember is Meta‘s 2026-11-10 small-tier entry in the Muse family — multimodal (text/image/audio/video), 1M-token context, $0.10 in / $0.50 out per Mtok. Pitched as the price-tier leader for the small-agent segment: The Decoder cites ~$0.14 per task on Artificial Analysis vs $2.39 for Muse Spark 1.3, and an AA Intelligence Index score of 36 (just below Spark 1.3’s 37). On FrontierCode 2.0-AA, Ember scores 27.2% vs Claude Haiku 5.5‘s 34.9% — a meaningful capability gap. The release is a price-tier play, not a Pareto-frontier winner; the useful practitioner read is that Ember belongs in the agent-execution routing fabric for cost-sensitive subtasks, not in the primary-reasoning slot.

Timeline

  • 2026-11-10-AI-Digest — Meta ships Muse Ember as the smallest entry in the Muse family — multimodal (text/image/audio/video), 1M-token context, $0.10 in / $0.50 out per Mtok. The Decoder cites ~$0.14 per task on Artificial Analysis vs $2.39 for Muse Spark 1.3, and an AA Intelligence Index score of 36 (just below Spark 1.3’s 37). On FrontierCode 2.0-AA, Ember scores 27.2% vs Claude Haiku 5.5‘s 34.9% — a meaningful capability gap. Reframe worth carrying: The Decoder’s “small AI built for the repetitive work agents actually do” framing is about the price-per-task tier (where Ember leads on the AA curve), not frontier displacement — Ember is a price-tier play, not a Pareto-frontier winner, and the Haiku 5.5 FrontierCode delta is the clearest capability-vs-price trade-off in the release. For practitioners, the useful read is “Ember belongs in the agent-execution routing fabric for cost-sensitive subtasks, not in the primary-reasoning slot”. First time Meta’s price-tier play has an explicit frontier-benchmark delta attached to it on release day — intelligence-per-dollar is now a published-at-launch metric for the cost-tier models, not a post-launch Artificial Analysis finding. Log against MOC - Open Source Models.

Key Developments

  1. Price-Tier Leader on the AA Curve, Not a Frontier Displacement (November 10, 2026): $0.10/$0.50 per Mtok with ~$0.14 per task on Artificial Analysis vs $2.39 for Muse Spark 1.3 — Ember leads the small-agent cost tier on the AA curve. AA Intelligence Index score of 36 sits just below Spark 1.3’s 37, which places Ember inside the family, not ahead of it; the design axis is per-task cost, not raw capability.

  2. Explicit Frontier-Benchmark Delta on Release Day — 27.2% vs Haiku 5.5’s 34.9% on FrontierCode 2.0-AA: First time Meta’s price-tier play has an explicit frontier-benchmark delta attached to it on release day. Claude Haiku 5.5 beating Ember by ~7.7pp on FrontierCode 2.0-AA while Ember wins on cost-per-task is the kind of disaggregated comparison the small-agent tier has been missing — carry as “intelligence-per-dollar is now a published-at-launch metric for the cost-tier models, not a post-launch Artificial Analysis finding”.

  3. Routing-Fabric Model, Not Primary-Reasoning Model: The practitioner read: Ember belongs in the agent-execution routing fabric for cost-sensitive subtasks, not in the primary-reasoning slot — same shape as Claude Haiku 5.5‘s role in Anthropic’s own routing posture, and the open question is whether downstream harnesses route to Ember on cost or stay on Haiku on the FrontierCode capability delta.

See also: Meta, Muse, Muse Spark, Muse Glimmer, Muse Image, Muse Code, Claude Haiku 5.5, MOC - Open Source Models.