Meta Muse Spark 1.3 Max matches Claude Fable 5 and GPT-5.6 Sol at 4-8x lower cost
AccBalanced · x · 2026-09-08
Per Vals AI's benchmark (shared by Scale AI CEO Alexandr Wang), Meta's Muse Spark 1.3 (Max) matches Claude Fable 5 and GPT-5.6 Sol on the Vals Index while costing 4x-8x less — a notable shift in frontier-model price/performance.
Related event: Meta Muse Spark 1.3 Max Matches Flagship Rivals at a Fraction of the Price(3 posts)→
More from Models
- Burning $7K/day in API costs: data analysis goes brrr — xeophon · 2026-09-08
- Data labs that outsource training give up all the value, argues X thread — joecole · 2026-09-08
- Local commercial license for MiniMax H3 costs $5,000/month, small studio says it's prohibitive — jonask86 · 2026-09-08
- speculation: OpenAI's Astra appears trained as a recursive language model — xeophon · 2026-09-08
- Same prompt, four models, one motorcycle: Blender MCP comparison crowns Astra — a10ondr · 2026-09-08
- Daily AI brief: Mistral raises €3B at €21B+, Meta Muse agent in closed alpha, DevDay Sept 29 — testingcatalog · 2026-09-08