TBC launches Mischief, a Wan 2.2-based image-to-video model, on fal.ai October 30
The Biological Computing Co. (TBC) is launching TBC Mischief, an image-to-video model based on Wan 2.2, through fal.ai on October 30. TBC says it generates video five times faster at 80% lower cost than the underlying model, while improving output quality.
The Biological Computing Co. (TBC) is launching TBC Mischief, an image-to-video model based on Wan 2.2, through fal.ai on October 30.
According to TBC, its version generates video five times faster at 80% lower cost than the underlying Wan 2.2 model, while also improving output quality. These are company claims, and no independent benchmark results have been published alongside them.
Neuron-derived approach
TBC describes the model as inspired by measurements of living neurons. The company's stated aim is to shorten the repeated-attempt loop that can come with generating a short AI video before a usable clip emerges.
Key details
- Model: TBC Mischief
- Type: Image-to-video
- Base model: Wan 2.2
- Availability: fal.ai, from October 30
- TBC's claims: five times faster generation, 80% lower cost, improved output quality, all relative to the underlying model
- Pricing, context window, parameter count and benchmark scores: not yet disclosed
What this means
Mischief is a derivative of Wan 2.2, and TBC's speed and cost figures are measured against that base model. Whether the improvements hold in practice will depend on testing by users on fal.ai, since the quality claim comes from TBC and no benchmark data accompanies it.
Related Articles
NVIDIA and MIT team releases Open-H3-Regenerate-2K, regenerating MiniMax H3 video drafts at up to native 4K
The Efficient-Large-Model team, made up of NVIDIA and MIT researchers, released Open-H3-Regenerate-2K weights on Hugging Face. Built on MiniMax H3, it regenerates low-resolution video drafts at 2560×1440 and up to native 3840×2160. The team reports a 3.9x speedup on a single B200 and a higher UniPercept score than the official H3-Regenerate-2K API.
Pika releases H3 Flex, an open-weight video model that generates synchronized audio in six sampling steps
Pika Labs has published Pika H3 Flex, a post-trained H3 model that generates video with synchronized audio in six sampling steps and accepts up to 12 reference inputs. Open weights run locally at native 480p and 768p, while Pika's hosted API adds 1080p, 2K and 4K output.
Microsoft releases Mage-VL-Cowork, a 4B computer-use agent scoring 66.0% on OSWorld-Verified
Microsoft has released Mage-VL-Cowork, a 4B-parameter vision-language policy that combines GUI control and terminal execution in one model. Microsoft claims it is the best 4B model on OSWorld-Verified, at 66.0% under a 50-step budget. Weights are published under Apache-2.0.
Samaya AI's Sterling scores 58.9% on FrontierFinance at $0.61 per question, ahead of Claude Fable 5.1
Samaya AI has released Sterling, a finance research agent model post-trained from Nemotron 3.5 Lightning with 3B active parameters. It scores 58.9% on FrontierFinance at $0.61 per question, against 55.9% at $6.89 for Claude Fable 5.1.
Comments
Loading...