Meta's Muse Spark 1.3 beats GPT-5.6 and Opus 5 on coding benchmarks
It led on long-context and coding scores, but still trailed Opus 5 on broad knowledge-work benchmarks
00
Tag
long-context
That's the last story.
Tag
It led on long-context and coding scores, but still trailed Opus 5 on broad knowledge-work benchmarks
That's the last story.