Pika launches H3 Flex: faster-than-real-time video model, up to 80% cheaper than H3Max
techNmak · x · 2026-11-01
Pika released Pika H3 Flex, a post-trained video model aimed at professional iteration speed:
- Speed: averages 1.8x faster than H3Max across matched t2v/i2v/r2v configs (DiT time); a 15-second clip can generate in as little as 7 seconds
- Price: up to 80% cheaper than H3Max
- Control: resolutions up to 4K, up to 12 reference inputs, precision editing via references and prompts, with a draft-at-768p / finish-at-4K workflow
- Audio: new exportable audio stems to isolate and refine any part of the soundscape
Available via the Pika API for creators who want to iterate at the speed of thought.
More from Models
- Robot Intelligence Index debuts: GPT-6 Astra tops robot model ranking at 1905 — haider1 · 2026-11-01
- Reddit rant: Google and OpenAI quietly nerfing $20-tier quotas by 30-50% — Outside-Clothes-8358 · 2026-11-01
- HuggingFace releases FineDNA dataset, turning Qwen3-14B into a DNA model that beats GPT-6.1 — loubnabnl · 2026-11-01
- Prediction: Local Models Will Be the Biggest AI Story of 2027, Disrupting Inference Business — ZeYanjie · 2026-11-01
- Carbon-X: midtrained Qwen3-14B beats every open LLM on DNA benchmarks — LoubnaBenAllal1 · 2026-11-01
- DeepSeek-V4.1-Flash uncensored GGUF quant hits Hugging Face trending — AtomicChat · 2026-11-01