Muse Spark 1.3 Improves Coding Efficiency; Launch Index Trails Fable 5.1
Meta Superintelligence Labs released Muse Spark 1.3 — about 20% fewer tool calls and 25% fewer tokens than 1.2 on internal coding. Artificial Analysis Intelligence 61 at xhigh versus Fable 5.1 max at 66. Fourth Spark in five months.
Meta Superintelligence Labs released Muse Spark 1.3 on September 2 for Muse Code and the Meta Model API. Meta’s internal coding comparisons report roughly 20% fewer tool calls and 25% fewer tokens than Spark 1.2.
Claims and independent checks#
Artificial Analysis’s dated September 2 evaluation describes the fourth Spark release in five months. It measured 61 on its Intelligence Index for xhigh, while max, then in limited partner preview, scored 62. Claude Fable 5.1 max scored 66 in that comparison.
Those index results are separate from Meta’s internal coding-efficiency tests. Artificial Analysis also measured $0.55 per index task for xhigh at $1.25/$4.25 per million input/output tokens; it said max pricing was not yet public.
Availability caveat#
Meta’s current launch page says max reasoning is available. The dated Artificial Analysis report establishes its limited-preview status at launch, but the inspected pages do not establish September 4 as the exact general-availability date. Meta lists an open-weight release on its roadmap without a date.
Correction (October 14, 2026): Removed unsupported launch-day executive quotes and the unverified September 4 max-availability date; distinguished the current company page from the dated launch evaluation.