--- title: Meta upgrades Muse Spark for long-horizon coding without... description: Meta released Muse Spark 1.3, an upgrade to its coding-focused AI model line, claiming roughly 25% fewer tokens and 20% fewer tool calls than its predecessor... canonical: https://daily.dev/posts/meta-upgrades-muse-spark-for-long-horizon-coding-without-raising-api-prices-dww2bbtnr twitter:card: summary_large_image twitter:site: @dailydotdev og:type: website og:site_name: daily.dev og:title: Meta upgrades Muse Spark for long-horizon coding without raising API prices | daily.dev og:description: Meta released Muse Spark 1.3, an upgrade to its coding-focused AI model line, claiming roughly 25% fewer tokens and 20% fewer tool calls than its predecessor... og:url: https://daily.dev/posts/meta-upgrades-muse-spark-for-long-horizon-coding-without-raising-api-prices-dww2bbtnr og:image: https://api.daily.dev/og/posts/dWW2Bbtnr.png og:image:alt: Meta upgrades Muse Spark for long-horizon coding without raising API prices og:image:width: 1200 og:image:height: 630 og:locale: en --- > ## Documentation Index > Fetch the complete documentation index at: https://daily.dev/llms.txt > Use this file to discover all available pages before exploring further. # Meta upgrades Muse Spark for long-horizon coding without raising API prices **[InfoWorld](https://daily.dev/sources/infoworld)** · 4 min read · 0 upvotes · 0 comments ## Summary Meta released Muse Spark 1.3, an upgrade to its coding-focused AI model line, claiming roughly 25% fewer tokens and 20% fewer tool calls than its predecessor while keeping the same per-token API pricing. Analysts say this could let development teams tackle more complex agentic coding tasks (large codebases, debugging, multi-file changes, testing) at similar cost, though the Jevons Paradox may offset overall savings as usage scales up. Independent benchmarks from Artificial Analysis and critiques from Eesel AI and Reddit users dispute Meta's efficiency claims, noting the comparison pits 1.3's max reasoning mode against 1.2's xhigh mode rather than a clean generational comparison, and that token usage varies significantly by workload. Despite the disputed benchmarks, analysts note Muse Spark 1.3 undercuts Anthropic and OpenAI flagship pricing by 4-to-6x while remaining competitive. ## Full article daily.dev links to this article rather than hosting it. Read it at the original source: ## Questions this post answers ### What improvements does Meta's Muse Spark 1.3 model bring for coding tasks compared to Muse Spark 1.2? Muse Spark 1.3 completes coding tasks using about 25% fewer tokens and 20% fewer tool calls than Muse Spark 1.2, according to Meta's internal benchmarks. It takes fewer turns when unnecessary, produces less verbose responses, and better sustains long-running tasks by generating its own context, correcting plan gaps, and following complex instructions. Per-token API pricing remains unchanged from the previous version. _daily.dev helps teams track model releases like Muse Spark 1.3 before committing budget to a new coding assistant._ ### Why are analysts skeptical of Meta's efficiency claims for Muse Spark 1.3? Meta's benchmark comparison pits Muse Spark 1.3's max reasoning mode, still gated behind safety testing at launch, against Muse Spark 1.2's xhigh mode, meaning part of the reported gain reflects a reasoning-tier change rather than a pure generational improvement. Independent testing from Artificial Analysis also found the model used about 100 million tokens on its Intelligence Index evaluation versus a 71 million-token median across other models, contradicting Meta's efficiency claims. _Developers comparing coding model benchmarks against real workloads can follow this debate on daily.dev._ ### How does Muse Spark 1.3's pricing compare to Anthropic and OpenAI's flagship coding models? Muse Spark 1.3 undercuts Anthropic and OpenAI's flagship model pricing by 4-to-6x while still producing results competitive with those more expensive models, according to industry analysts. This pricing gap, combined with unchanged per-token costs from the prior Muse Spark version, positions it as a lower-cost alternative for enterprises running large-scale software development workloads. _Teams weighing coding model costs against Anthropic and OpenAI pricing can track these comparisons on daily.dev._ ## Similar posts on daily.dev - [Meta debuts Muse Spark 1.1 and it isn’t free](https://daily.dev/posts/meta-debuts-muse-spark-1-1-and-it-isn-t-free-h3ju1gqtf) · The New Stack · 6 upvotes · 0 comments --- Tags: [#ai](https://daily.dev/tags/ai), [#ai-agents](https://daily.dev/tags/ai-agents) [View this post on daily.dev](https://daily.dev/posts/meta-upgrades-muse-spark-for-long-horizon-coding-without-raising-api-prices-dww2bbtnr) ```json {"@context":"https://schema.org","@graph":[{"@type":"Organization","@id":"https://daily.dev/#organization","name":"daily.dev","url":"https://daily.dev","logo":{"@type":"ImageObject","url":"https://daily.dev/apple-touch-icon.png","width":180,"height":180},"sameAs":["https://twitter.com/dailydotdev","https://github.com/dailydotdev","https://www.linkedin.com/company/daily-dev-ltd"]},{"@type":"WebSite","@id":"https://daily.dev/#website","url":"https://daily.dev","name":"daily.dev","publisher":{"@id":"https://daily.dev/#organization"},"potentialAction":{"@type":"SearchAction","target":{"@type":"EntryPoint","urlTemplate":"https://daily.dev/search?q={search_term_string}"},"query-input":"required name=search_term_string"}}]} {"@context":"https://schema.org","@type":"TechArticle","headline":"Meta upgrades Muse Spark for long-horizon coding without raising API prices","url":"https://daily.dev/posts/meta-upgrades-muse-spark-for-long-horizon-coding-without-raising-api-prices-dww2bbtnr","mainEntityOfPage":{"@type":"WebPage","@id":"https://daily.dev/posts/meta-upgrades-muse-spark-for-long-horizon-coding-without-raising-api-prices-dww2bbtnr"},"datePublished":"2026-09-03T12:09:01.093Z","dateModified":"2026-09-04T15:32:56.683Z","description":"Meta released Muse Spark 1.3, an upgrade to its coding-focused AI model line, claiming roughly 25% fewer tokens and 20% fewer tool calls than its predecessor...","image":"https://media.daily.dev/image/upload/f_auto,q_auto/v1/posts/d43c94b71961033df084a5c8ba2ff54d?_a=AQAEuop","thumbnailUrl":"https://media.daily.dev/image/upload/f_auto,q_auto/v1/posts/d43c94b71961033df084a5c8ba2ff54d?_a=AQAEuop","isAccessibleForFree":true,"articleSection":"InfoWorld","inLanguage":"en","publisher":{"@type":"Organization","name":"daily.dev","url":"https://daily.dev","logo":{"@type":"ImageObject","url":"https://daily.dev/apple-touch-icon.png","width":180,"height":180}},"author":{"@type":"Organization","name":"InfoWorld","logo":"https://media.daily.dev/image/upload/t_logo,f_auto/v1/logos/bf6d68a999064029b0bb09aa6268f1f3","url":"https://daily.dev/sources/infoworld"},"commentCount":0,"discussionUrl":"https://daily.dev/posts/meta-upgrades-muse-spark-for-long-horizon-coding-without-raising-api-prices-dww2bbtnr","interactionStatistic":[{"@type":"InteractionCounter","interactionType":{"@type":"LikeAction"},"userInteractionCount":0},{"@type":"InteractionCounter","interactionType":{"@type":"CommentAction"},"userInteractionCount":0}],"keywords":"ai,ai-agents","timeRequired":"PT4M"} {"@context":"https://schema.org","@type":"BreadcrumbList","itemListElement":[{"@type":"ListItem","position":1,"name":"Home","item":"https://daily.dev"},{"@type":"ListItem","position":2,"name":"InfoWorld","item":"https://daily.dev/sources/infoworld"},{"@type":"ListItem","position":3,"name":"Meta upgrades Muse Spark for long-horizon coding without raising API prices"}]} {"@context":"https://schema.org","@type":"FAQPage","@id":"https://daily.dev/posts/meta-upgrades-muse-spark-for-long-horizon-coding-without-raising-api-prices-dww2bbtnr#faq","mainEntity":[{"@type":"Question","name":"What improvements does Meta's Muse Spark 1.3 model bring for coding tasks compared to Muse Spark 1.2?","acceptedAnswer":{"@type":"Answer","text":"Muse Spark 1.3 completes coding tasks using about 25% fewer tokens and 20% fewer tool calls than Muse Spark 1.2, according to Meta's internal benchmarks. It takes fewer turns when unnecessary, produces less verbose responses, and better sustains long-running tasks by generating its own context, correcting plan gaps, and following complex instructions. Per-token API pricing remains unchanged from the previous version. daily.dev helps teams track model releases like Muse Spark 1.3 before committing budget to a new coding assistant."}},{"@type":"Question","name":"Why are analysts skeptical of Meta's efficiency claims for Muse Spark 1.3?","acceptedAnswer":{"@type":"Answer","text":"Meta's benchmark comparison pits Muse Spark 1.3's max reasoning mode, still gated behind safety testing at launch, against Muse Spark 1.2's xhigh mode, meaning part of the reported gain reflects a reasoning-tier change rather than a pure generational improvement. Independent testing from Artificial Analysis also found the model used about 100 million tokens on its Intelligence Index evaluation versus a 71 million-token median across other models, contradicting Meta's efficiency claims. Developers comparing coding model benchmarks against real workloads can follow this debate on daily.dev."}},{"@type":"Question","name":"How does Muse Spark 1.3's pricing compare to Anthropic and OpenAI's flagship coding models?","acceptedAnswer":{"@type":"Answer","text":"Muse Spark 1.3 undercuts Anthropic and OpenAI's flagship model pricing by 4-to-6x while still producing results competitive with those more expensive models, according to industry analysts. This pricing gap, combined with unchanged per-token costs from the prior Muse Spark version, positions it as a lower-cost alternative for enterprises running large-scale software development workloads. Teams weighing coding model costs against Anthropic and OpenAI pricing can track these comparisons on daily.dev."}}]} ```