← Back to live feed1 story
Meta's latest artificial intelligence model surpassed DeepSeek to capture the most daily users on a prominent tracking list. The Muse Spark 1.3 (xHigh) variant scored 1,623 points in Code Arena: WebDev, placing the company 10th overall on the web development coding leaderboard. This ranking marks a significant climb from the 30th position held by the previous Muse Spark 1.2 version.
Performance for the new model is roughly equal to Claude Fable 5 and GPT-5.6 Sol. Current benchmark scores rely on an AutoEval reward model trained on human preference data and will be updated as live human votes are processed. Internal comparisons indicate a 25% reduction in token usage and 20% fewer tool calls than the predecessor.
Sign in to suggest edits
Key sources
- SOURCE@arena“Muse Spark 1.3 (xHigh) is a huge improvement from previous models Muse Spark 1.1 at #25 (1540 pts), and Muse Spark 1.2 (xHigh) at #30 (1534 pts)”x.com
- SUPPORT@thdxr“meta muse spark has dethroned deepseek as the most used model of the day”x.com
- SOURCE@jilin_14“The leap from 1.0 to 1.3 is nothing short of a miracle”x.com
- SUPPORT@wesroth“from 55.0 to 75.4 on DeepSWE.”x.com
- SUPPORT@wesroth“And the limited-preview max version pushes that score to 62.”x.com
- SUPPORT@kimmonismus“Meta Muse Spark 1.3 is now free on OpenCode”x.com
- SUPPORT@eliebakouch“muse spark 1.3 doesn't seem to be better on other long context benchmarks => very likely benchmaxed on this specific task”x.com