---
format: "aidr-story-markdown/v1"
id: "8cca74416a942173912aba303622a58545d9f11a5424780dd3802fe404ee08f7"
canonical_url: "https://aidr.today/8cca7441?lang=en"
title: "Meta Rolls Out Muse Spark 1.3 at $1.72 Per Task, Lowest Cost Coding Agent Above 60 Score"
lang: "en"
requested_lang: "en"
available_langs: ["en","vi"]
translation_fallback: null
fallback_fields: []
published_at: "2026-09-03T08:49:06.000Z"
category: "Models"
topics: ["meta","agent","coding","open-source"]
source_urls: ["https://huggingnews.com/ai/meta-rolls-out-muse-spark-13-at-172-per-task-lowest-cost-coding-agent-ab-2e3a768c","https://x.com/shishirpatil_/status/2095307695050178810","https://x.com/ArthurMacwaters/status/2095315716442595748","https://x.com/ArtificialAnlys/status/2095317750839476393","https://x.com/ArtificialAnlys/status/2095317755222503459","https://x.com/WesRoth/status/2095368689763619309","https://x.com/alexandr_wang/status/2095308526575100127","https://x.com/finkd/status/2095232032896946311"]
summary: "The new coding and agentic model is available now in Muse Code and the Meta Model API, and Meta said it is its biggest jump yet in coding and agentic work. Artificial Analysis said the xhigh version scored 64 on its Coding Agent Index at $1.72 per task, the lowest cost of any agent above a 60 score. A limited-preview max version for Meta's partners scored 68 and ranked second behind Claude Opus 5. Meta said Muse Spark 1.3 uses about 20% fewer tool calls and about 25% fewer tokens than Muse Spark 1.2 in internal comparisons. Separate benchmark results put DeepSWE at 75.4 versus 55.0 for Muse Spark 1.2 and Terminal-bench 2.1 at 88.8, matching GPT-5.6 Sol. Meta said max reasoning and a Muse Spark open weights release are coming later."
---

# Meta Rolls Out Muse Spark 1\.3 at $1\.72 Per Task, Lowest Cost Coding Agent Above 60 Score

> [Open the canonical story](<https://aidr.today/8cca7441?lang=en>)

**Published:** 2026-09-03T08:49:06.000Z
**Category:** Models
**Topics:** meta, agent, coding, open\-source

## Summary

The new coding and agentic model is available now in Muse Code and the Meta Model API, and Meta said it is its biggest jump yet in coding and agentic work\. Artificial Analysis said the xhigh version scored 64 on its Coding Agent Index at $1\.72 per task, the lowest cost of any agent above a 60 score\. A limited\-preview max version for Meta's partners scored 68 and ranked second behind Claude Opus 5\. Meta said Muse Spark 1\.3 uses about 20% fewer tool calls and about 25% fewer tokens than Muse Spark 1\.2 in internal comparisons\. Separate benchmark results put DeepSWE at 75\.4 versus 55\.0 for Muse Spark 1\.2 and Terminal\-bench 2\.1 at 88\.8, matching GPT\-5\.6 Sol\. Meta said max reasoning and a Muse Spark open weights release are coming later\.

## Sources

- [Story source](<https://huggingnews.com/ai/meta-rolls-out-muse-spark-13-at-172-per-task-lowest-cost-coding-agent-ab-2e3a768c>)
- [Story source](<https://x.com/shishirpatil_/status/2095307695050178810>)
- [Supporting source](<https://x.com/ArthurMacwaters/status/2095315716442595748>)
- [Supporting source](<https://x.com/ArtificialAnlys/status/2095317750839476393>)
- [Supporting source](<https://x.com/ArtificialAnlys/status/2095317755222503459>)
- [Supporting source](<https://x.com/WesRoth/status/2095368689763619309>)
- [Supporting source](<https://x.com/alexandr_wang/status/2095308526575100127>)
- [Story source](<https://x.com/finkd/status/2095232032896946311>)