MiniCPM-V 4.5 beats much larger models by focusing on efficiency instead of brute force. Instead, it combines (1) extreme video compression, (2) smarter OCR/document pretraining, (3) controllable fast/deep reasoning, and (4) efficient alignment. So an 8B model can outperform
MULTIMODAL AI
-
MiniCPM-V 4.5 Open Source Multimodal LLM Outperforms GPT-4o
By
–
GPT-4o level multimodal LLM running on your phone.
— Shubham Saboo (@Saboo_Shubham_) 5 septembre 2025
MiniCPM-V 4.5 outperforms GPT-4o, Gemini-2.0 Pro, and Qwen2.5-VL 72B on vision and language AI tasks.
It can even understand videos and perform OCR in 30+ languages.
And it's 100% Opensource. pic.twitter.com/wY70HlmzDSGPT-4o level multimodal LLM running on your phone. MiniCPM-V 4.5 outperforms GPT-4o, Gemini-2.0 Pro, and Qwen2.5-VL 72B on vision and language AI tasks. It can even understand videos and perform OCR in 30+ languages. And it's 100% Opensource.
-
Style Generation Through High-Dimensional Continuous Space Sampling
By
–
these styles are computed via random sampling of a continuous high dimensional space that is meant to contain all possible concepts of style, certainly there's stuff outside of this but it's a very powerful concept!
-
Planning with Reasoning using Vision Language World Model
By
–
Planning with Reasoning using Vision Language World Model pic.twitter.com/RZF8SVpV17
— AK (@_akhaliq) 4 septembre 2025Planning with Reasoning using Vision Language World Model
-
AI Self-Portrait: Machine Emotional Expression Through Art
By
–
@NanoBanana Draw a self-portrait that reflects your emotions.
-
Sequential Art Comic Panel Generation with AI
By
–
2/5 Sequential art (comic panel / storyboard) A single comic book panel in a [art style] style. In the foreground, [character description and action]. In the background, [setting details]. The panel has a [dialogue/caption box] with the text "[Text]". The lighting creates a
-
Perplexity Comet Mobile App Available for Pre-Order Android
By
–
Comet is coming soon to mobile and is now available for pre-orders on Android Play Store https://
play.google.com/store/apps/det
ails?id=ai.perplexity.comet&hl=en_US
… -
Vision Language Models on Edge Devices: Practical Viability Assessment
By
–
📢New Post Alert:
— Satya Mallick (@LearnOpenCV) 4 septembre 2025
📙VLM on Edge Devices: Worth the Hype or Just a Novelty?
The rise of Vision Language Models (VLMs) has been meteoric but can they really run effectively on edge devices?
Our latest post is first in the series of experiments that we will continue for VLMs on… pic.twitter.com/qBuvMtS8AuNew Post Alert: VLM on Edge Devices: Worth the Hype or Just a Novelty? The rise of Vision Language Models (VLMs) has been meteoric but can they really run effectively on edge devices?
Our latest post is first in the series of experiments that we will continue for VLMs on -
Real-time Video Beta Invites Launch Today
By
–
real-time video is here.
— KREA AI (@krea_ai) 4 septembre 2025
today we start rolling out beta invites.
who wants to try? pic.twitter.com/evUG1Woy26real-time video is here. today we start rolling out beta invites. who wants to try?
-
Gen:48 Aleph Edition Winning Films Announced on Runway ML
By
–
Congratulations again to everyone who participated in Gen:48 Aleph Edition. Watch all of the winning films at http://
runwayml.com/gen48 (8/8)