Meta Rises 4% as Muse Spark 1.3 Claims Parity With Anthropic and OpenAI
📈 META shares jumped 4% to $614.57 after claiming Muse Spark 1.3 matches Anthropic and OpenAI in AI model performance.
🤖 CEO Mark Zuckerberg reported a 60% increase in daily user interactions with the Meta AI assistant since rebuilding it around the Muse Spark family.
💼 Business agents on WhatsApp and Messenger reached more than 1 million businesses per week, indicating strong enterprise adoption.
⚠️ Q2 2026 diluted EPS of $6.18 missed the $7.22 estimate due to legal and severance charges, ending a six-quarter beat streak.
🔮 The company plans an open weights release for Muse Spark models in the future while maintaining closed ecosystem options.
📅 Meta's next major catalyst is the September 23 Connect event, which will demonstrate Muse Spark 1.3 integration with new AI glasses.
📊 Alphabet (GOOGL) leads the sector with a 60% one-year return and 950 million Gemini users, creating a significant competitive gap for Meta to close.
🎯 Alexandr Wang, Meta's AI chief, described Muse Spark 1.3 as the company's biggest jump in model performance so far.
📉 Prediction markets give only 48% odds that Meta scores 55% on Humanity's Last Exam this year, casting doubt on the parity claim.
🛠️ Muse Spark 1.3 is rolling out via Muse Code and the Meta Model API with a max reasoning mode pending additional safety testing.
- Stock surged 4% to $614.57 on claims that Muse Spark 1.3 achieves AI model parity with top rivals Anthropic and OpenAI.
- Daily interactions with the Meta AI assistant increased by 60% since the company rebuilt its platform around the Muse Spark family.
- Business agents on WhatsApp and Messenger are engaging over 1 million businesses per week, demonstrating rapid enterprise traction.
- The company is rolling out Muse Spark 1.3 via Muse Code and the Meta Model API to developers immediately.
- Q2 2026 diluted EPS of $6.18 missed the $7.22 estimate due to legal and severance charges, ending a six-quarter beat streak.
- The stock rallied despite no independent benchmark confirmation yet for the claimed performance parity with competitors.
- Prediction markets assign only 48% probability that Meta scores 55% on Humanity's Last Exam this year, indicating skepticism about the model's capabilities.