Pika Labs launches Pika 2.5 with enhanced physics and sound effects

Pika Labs, a company founded by two Stanford graduates, has released its flagship video generation engine, Pika 2.5. This new version, available by default on all plans in 2026, boasts improved physics simulation, allowing for realistic object interactions like breaking glass. While Pika 2.5 includes synchronized sound effects, it does not generate music or dialogue. For those capabilities, Pika Labs offers Pikaformance, a separate model that creates talking or singing videos from photos and audio, with a focus on lip-sync and facial micro-expressions. AI

IMPACT Enhances realism in AI video generation with improved physics and synchronized sound effects.

RANK_REASON Frontier-lab model release with system card [lever_c_demoted from frontier_release: ic=1 ai=1.0]

Read on dev.to — LLM tag →

• Adam D'Angelo
• Greycroft
• Jared Leto
• Lightspeed
• Mellis, Inc.
• Nat Friedman
• Pika 2.5
• Pikaformance
• Pika Labs
• Runway Gen-4 Turbo
• Spark Capital
• Stanford

• model release
• product

AI-generated summary · Google Gemini · from 1 sources. How we write summaries →