Archive — US — techcrunch
Meta’s next Llama models may have upgraded voice features
Meta is set to release Llama 4, a new AI model with enhanced voice capabilities, which will allow users to interrupt mid-speech. The launch is anticipated soon as competition with models like DeepSeek intensifies, pressuring Meta to innovate in response to market demands.
Key points
- Meta plans to launch Llama 4, with upgraded voice features.
- The model will allow users to interrupt mid-speech, similar to OpenAI and Google offerings.
- Chris Cox describes Llama 4 as an 'omni' model, capable of handling speech, text, and other data.
- Competition from DeepSeek, which has produced better-performing models, has accelerated Llama's development.
- Meta has established war rooms to analyze DeepSeek's cost-effective model deployment strategies.
Why it matters
- The trend towards AI models incorporating voice features aligns with increasing consumer demand for interactive AI.
- Voice-enabled AI is becoming essential in enhancing user experiences across various platforms.
- Market competition drives innovation, reflecting historical patterns in tech development, reminiscent of early responses to groundbreaking technologies like smartphones.
Meta's Llama 4 represents a significant shift towards integrated voice capabilities in AI, emphasizing the urgent need for tech companies to innovate amidst fierce competition, with implications for future AI development through 2025.