Muse Spark 1.2 Contributor

Meta released 2026-08-21 proprietary

$0.125 per million tokens, balanced

No independent quality score.

No benchmark we track has measured this model. That is not the same as measuring it and finding it wanting — we simply cannot rank it, so we do not.

Our take

editorial — not a measurement

A deliberately cheap tier of Muse Spark 1.2 at $0.10/$0.20 with the same 1M context and full multimodal input, plus a $0.002 cache-read rate that is the lowest in the dataset. If the capability is anywhere near the flagship's, the price-per-capability is extraordinary. Meta describes it as aimed at developers starting out at low cost, which implies throughput or priority constraints that are not documented anywhere reachable. Worth a serious evaluation precisely because the published numbers look too good.

Strengths

  • $0.10/$0.20 with a 1M context window
  • $0.002 cache reads — lowest in the dataset
  • Full multimodal input including audio and video
  • Same nominal architecture as the 1498-Elo flagship

Weaknesses

  • Undocumented constraints implied by the 'contributor' framing
  • No benchmark score of its own
  • No first-party pricing page
  • Relationship to flagship quality is unverified

Reach for it when

  • Cost-floor multimodal experimentation
  • High-volume cache-heavy pipelines

Avoid it if

  • You need documented SLAs or throughput guarantees
  • Production work where undocumented limits are unacceptable

Sources: openrouter.ai

Every price dimension

Input$0.100/M
Output$0.200/M
Cached input$0.0020/M
Web search$0.0025/call

Independent scores

No independent benchmark has measured this model.

Capability

Context window1.0M
Input modestext, image, video, file, audio
Tool useyes
Reasoningalways on
Open weightsno

Provenance

Price sourceopenrouter.ai
Fetched2026-08-24
Quality dataunrated
Cross-checkedvendor page

Muse Spark 1.2 contributor tier is a reasoning model from Meta designed for developers who want to start building at an even lower cost. It’s meaningfully cheaper than Muse Spark...

Markdown for LLMs