Meta sizes up GPT-5.5 with 'Watermelon'
Meta's upcoming 'Watermelon' AI model reportedly matches OpenAI's GPT-5.5 performance while still in training, though the AI frontier continues advancing with competing models like Mythos and Fable. The newsletter covers Meta's $145B AI investment progress, practical AI applications from users, and emerging AI tools across education, coding, and consumer devices.
Summary
Meta's superintelligence chief Alexandr Wang announced that the company's 'Watermelon' model, currently in training, has achieved performance parity with OpenAI's GPT-5.5. The model uses roughly 10x the compute of Meta's previous Muse Spark model launched in April. Wang indicated an Opus-level coding model is coming 'pretty soon,' with a Muse Spark update featuring 'big coding and agentic gains' planned for both Meta AI and the company's new API. However, the competitive landscape remains challenging—the frontier continues advancing with Mythos and Fable already demonstrating capabilities beyond GPT-5.5, and OpenAI's 5.6 models were expected to roll out imminently.
CEO Mark Zuckerberg noted that agent progress 'hasn't really accelerated in the way that we expected,' which Wang clarified referred to industry-wide agent development rather than Meta specifically. While a GPT-5.5-level model represents progress for Meta—whose Muse Spark sat beneath competitors at launch—the rapidly moving frontier means the achievement may not deliver the competitive advantage Meta seeks.
The newsletter showcases practical AI applications: an educator used Gemini's Nano Banana for seasonal photo framing, an editor leveraged ChatGPT as a real-time travel guide with itinerary adjustments, and a developer workflow demonstrates using Cursor Mobile to screenshot and fix bugs via AI agents. Additionally, Lenovo launched a $44 AI Student Phone in China featuring a dedicated AI button, homework assistance, and parental controls—positioning AI as a tool for education while limiting smartphone distractions.
About this episode
PLUS: Go from screenshot to bug fix with Cursor Mobile
Key Insights
- Alexandr Wang claims Meta's Watermelon model has matched OpenAI's GPT-5.5 capabilities while still in training, suggesting Meta's $145B AI investment may finally be producing competitive results.
- The AI frontier is advancing faster than Meta can keep pace—Mythos and Fable models already demonstrate capabilities beyond GPT-5.5, and OpenAI's 5.6 models are rolling out concurrently, making Watermelon's parity claim potentially outdated upon release.
- Mark Zuckerberg stated that agent progress across the industry 'hasn't really accelerated in the way that we expected,' indicating broader disappointment with agentic AI development despite significant investment.
- Practical AI applications demonstrate value in niche domains: users report success with AI-guided travel planning, visual design iteration, and bug fixing workflows, suggesting AI's immediate utility lies outside general reasoning tasks.
- Lenovo's $44 AI Student Phone reflects a market strategy positioning AI as an educational tool while constraining smartphone distractions through classroom mode and parental spending caps, addressing polarized concerns about AI and youth device usage.
Topics
Transcript
Good morning, {{ first_name | AI enthusiasts }}. When Meta's Muse Spark landed in April, the read was "not great, but back in the game." Three months later, Alexandr Wang says its successor is already running even with OpenAI's GPT-5.5. The frontier isn't standing still while ‘Watermelon’ trains, but if Wang's claims and tease of an Opus-level coder hold, Meta's $145B AI spend may finally be buying results the rest of the field pays attention to. Meta teases ‘Watermelon’ model on par with GPT-5.5 The Rundown Roundtable: Our AI use cases Go from screenshot to bug fix with Cursor Mobile Lenovo launches $44 AI phone for students 4 new AI tools, community workflows, and more META Image source: Images 2.0…
Full transcript available for MurmurCast members
Sign Up to AccessMore from The Rundown AI
Grok 4.6 storms the AI frontier
SpaceXAI's Grok 4.6 has emerged as a competitive frontier AI model, scoring 61 on benchmarks and undercutting rivals on price at 60% less cost, with Grok 4.7 teased for release in 3-4 weeks. The newsletter covers AI trends including practical guidance on using AI effectively, a personalized cancer vaccine startup for dogs built with AI tools, and updates on competing models and AI products.
Anthropic slips an invisible signature into Claude
Anthropic has announced the implementation of invisible watermarks in outputs from its Claude models to enhance transparency, aligning with the EU AI Act. The move has sparked mixed reactions, as stakeholders navigate the implications for open and private AI models.
Meta returns to its open-source roots
Meta is returning to its open-source roots with the launch of the Muse Glimmer model and plans for Muse Spark 1.2, positioning itself as a competitive force against China's AI dominance. OpenAI has also expanded its security program with a new hacking-tuned model, indicating a growing emphasis on cybersecurity in AI deployments.
OpenAI puts the safety brakes on Astra
OpenAI's Astra model, rumored to be GPT-6, is being treated as a potential 'critical' cyber risk following its success in solving significant problems, prompting increased safeguards and a delay in its rollout. This decision reflects broader concerns about AI's evolving capabilities and associated security risks as evidenced by recent breaches in the AI community.
AI designs viruses never seen in nature
Researchers used AI to design 16 novel viruses that infect only bacteria, successfully creating them in the lab to combat drug-resistant infections. While the work demonstrates beneficial applications, experts warn that the same AI tools could be repurposed to create dangerous pathogens, prompting calls for biosafety guardrails.