Muse Spark 1.3
Meta's leaner agentic-coding model — the fourth Muse Spark in five months, keeping the 1M-token context window but using ~20% fewer tool calls and ~25% fewer tokens than 1.2.
Last verified · 2026-10-23 · by Moe Ameen
Published 2026-09-02 · Updated
Short answer
Muse Spark 1.3 is Meta's agentic coding model, released September 2, 2026 and available in Muse Code and the Meta Model API. Meta says it uses about 20% fewer tool calls and 25% fewer tokens than version 1.2, it keeps a 1 million-token context window, and standard API pricing is $1.25 per million input tokens and $4.25 per million output. It returns text and code, so publishing still takes another tool.
Key takeaways
- Muse Spark 1.3 launched September 2, 2026 for long, multi-step coding and agent work.
- Meta reports about 20% fewer tool calls and 25% fewer tokens than Muse Spark 1.2.
- Standard API pricing is $1.25 per million input tokens and $4.25 per million output.
- It reads text, images and video, but it only writes text and code back.
- To publish what it drafts, you still need a tool that renders and schedules posts.
What is Meta Muse Spark 1.3?
Muse Spark 1.3 is Meta Superintelligence Labs' agentic-coding model, launched September 2, 2026 and announced by CEO Mark Zuckerberg. It is the fourth Muse Spark model released in about five months, and unlike the earlier jumps, 1.3 is an efficiency-and-reliability refinement rather than a ground-up overhaul.
The design center is long-running, multi-step agentic work, not one-shot chat. Muse Spark 1.3 keeps the 1 million-token context window and is tuned to sustain longer-horizon tasks — collaborating with a user and juggling several workflows in a single long thread, following complex instructions more reliably, and orchestrating tools with less waste.
The headline improvement over version 1.2 is doing more with less: Meta reports roughly 20% fewer tool calls and about 25% fewer tokens on comparable tasks, plus a cleaner, less verbose coding style. On safety it claims stronger resistance to adversarial inputs and prompt injection and better calibration on what counts as an irreversible action during agentic tasks. Meta's announcement page now says Muse Spark 1.3 with max reasoning is available on Muse Code and the Meta Model API.
The multimodal side is where it touches creative work. Muse Spark 1.3 reads text, images, and video and reasons over them, which is useful for summarizing a long recording or describing a clip in detail. You reach it through Muse Code (Meta's coding agent) and the Meta Model API, and it is also listed on OpenRouter.
Standard API pricing held at $1.25 per million input tokens and $4.25 per million output (cached input $0.15 per million), with a lower-cost "contributor" variant whose traffic Meta uses to improve its products. VKTR reported that the max configuration of 1.3 scored 62 on the Artificial Analysis Intelligence Index, behind only two Claude models.
Muse Spark 1.3 is a reasoning-and-coding model, not a content-production tool. It can draft copy, write a script, summarize a video, or reason over a document, but it hands back raw text or code in an agent session or an API response. It does not keep a brand voice across a week of posts, render a carousel or an avatar video, size anything per platform, or schedule and publish to your social channels. Treat the specific benchmarks, the availability of max reasoning, and the exact pricing as a launch-window snapshot — verify current numbers on Meta's own pages.
What can you make with Muse Spark 1.3?
- Drafted copy, scripts, hooks, and outlines from a prompt in Muse Code or via the Meta Model API
- Structured summaries and key-moment lists from a long recording or transcript you feed it (multimodal input)
- Detailed descriptions and captions of an image or a video (multimodal reading, not creation)
- Agentic, multi-step coding tasks — bug fixes, feature builds, and migrations — with fewer tool calls than 1.2
- Long-horizon automations that hold context across several workflows in a single thread
- Custom apps and tools built on the Meta Model API, including through OpenRouter
What is the most useful part of Muse Spark 1.3 for creators?
The most creator-relevant thing about Muse Spark 1.3 is its multimodal reading: hand it a 40-minute webinar recording, or a podcast transcript and it will pull out the structure — the key moments, the quotable lines, the section outline. That is genuinely useful raw material. But it stops at a text description of your video; it cannot touch the video itself.
It will tell you "the strong segment starts around minute 12," but it will not cut that segment into a vertical clip, reframe it to keep the speaker centered, burn in animated captions, or push it to TikTok. That gap — between knowing what to make and actually rendering and shipping it — is exactly the span Kompozy covers.
Here is the concrete handoff. Take the outline or key-moments list Muse Spark 1.3 produced from your long-form source and bring it into Kompozy. Kompozy's Clipped Shorts turn the actual long-form video into vertical cuts; the same brief fans into a brand-exact Carousel, Quote Graphics built from the lines the model surfaced, a Blog Article and an Email Newsletter, and native Text Posts — all held to one voice by your Persona Brief and banned-word filters, and each rendered as finished media, not a description of media.
Then Kompozy schedules and publishes the whole set across eight social platforms plus blog and email from one queue with Autopilot and a per-post review pipeline. Muse Spark 1.3 reads your source and tells you what is in it; Kompozy is what turns that reading into the finished, on-brand posts and gets them live.
- Feed your long-form source — a recording or transcript — to Muse Spark 1.3 in Muse Code or via the API, and ask for an outline plus the strongest moments and quotable lines.
- Bring that outline (and the original video, if you have one) into Kompozy as the seed for a content unit.
- Use Clipped Shorts to cut the long-form video into vertical clips, and turn the outline into a Carousel, Blog Article, and Email Newsletter, all rendered on-brand.
- Let Kompozy build Quote Graphics from the surfaced lines and fan the idea into native Text Posts, held to one voice by your Persona Brief.
- Schedule and publish the full set across TikTok, Reels, Shorts, YouTube, LinkedIn, X, Pinterest, and Threads from one queue with Autopilot.
What does this look like in Kompozy?
A real quote graphic made in the Kompozy workspace. It shows the format a pulled line from a long recording can become; it was not made from Muse Spark output.
Who is this for, and who should skip it?
Best for:
- Developers running long agentic coding tasks who want fewer tool calls per job
- Teams that need a model to read a long recording or transcript and pull out its structure
- Builders already on OpenRouter or the Meta Model API who want a cheaper frontier option
Not ideal for:
- Creators who want finished videos, carousels or graphics; it only returns text
- Anyone who needs a fixed brand voice across a week of posts without another tool
- Teams that cannot let Meta use their traffic, if they were counting on the cheaper contributor endpoint
Frequently asked questions
What is Meta Muse Spark 1.3?
Muse Spark 1.3 is Meta Superintelligence Labs' agentic-coding model, launched September 2, 2026 as the fourth Muse Spark release in about five months. It keeps the 1 million-token context window and refines version 1.2 for efficiency — roughly 20% fewer tool calls and 25% fewer tokens on comparable tasks — and is available in Muse Code and on the Meta Model API.
What is new in Muse Spark 1.3 versus 1.2?
Meta reports about 20% fewer tool calls and 25% fewer tokens on comparable tasks, plus a cleaner coding style, better instruction-following on long tasks, stronger prompt-injection resistance, and better calibration on irreversible actions. Meta's announcement page now says the max reasoning setting is available on Muse Code and the Meta Model API.
How much does the Muse Spark 1.3 API cost?
On the Meta Model API, the standard model is $1.25 per million input tokens and $4.25 per million output (cached input $0.15 per million), unchanged from 1.1 — but 1.3 uses about a quarter fewer tokens per task, so the effective cost is lower. A cheaper "contributor" variant is also offered in exchange for Meta using your traffic to improve its products. Confirm current numbers on Meta's pricing page.
Can Muse Spark 1.3 make and publish social content?
No. Muse Spark 1.3 drafts text and code and can summarize or describe media you feed it, but it returns raw output in an agent session or API response. It does not render video or carousels, keep a brand voice across a week, size posts per platform, or schedule and publish anywhere. A content engine like Kompozy turns its drafts into finished, on-brand formats and publishes them across platforms.
How do we know this?
We checked every figure, date, price and quote on this page against the sources listed below on 2026-10-23, and removed or corrected anything we could not confirm. Benchmark coverage comes from VKTR and pricing from DataNorth and eesel, because Meta's launch post does not list them. Kompozy plan and format details come from kompozy.io/pricing.
Sources
- Meta AI Research — Introducing Muse Spark 1.3 (read 2026-10-23)
- VKTR — Meta's Muse Spark 1.3 posts frontier coding scores at a fraction of rival pricing (read 2026-10-23)
- DataNorth — Meta launches Muse Spark 1.3 (read 2026-10-23)
- eesel AI — Muse Spark 1.3 pricing (read 2026-10-23)
- OpenRouter — Muse Spark 1.3 (read 2026-10-23)
- Kompozy — Pricing (read 2026-10-23)
Related tools
- Muse Spark 1.1 — Meta's multimodal reasoning model built for agentic coding and computer use — with a 1M-token context window and parallel sub-agents, now open to developers on the Meta Model API.
- Muse Image — Meta's first in-house AI image model — you can @-mention a public Instagram account and it pulls that person's public photos into the generated image.
- GPT-5.6 — OpenAI's three-tier frontier model family — Sol, Terra, and Luna — with sharper image reading and stronger text-and-interface generation.
- Claude Sonnet 5 — Anthropic's cheaper, more agentic mid-tier Claude model — close to Opus 4.8 performance at a fraction of the price.