Limited-time offer: 20% off PixVerse V6 Max. Ends November 30. Try in Playground
Browse our latest posts to learn more about us and what our platform has to offer.
Guides
We gave Qwen Audio 3.0 ASR, Qwen Audio 3.1 ASR, StepAudio 3 ASR Max and Deepgram Nova-3 the same voicemail with a name, an invoice number, a price, a date and a phone number. Read all four.
We asked the four Qwen Audio realtime models the same spoken question over a WebSocket, one shot each. Hear every reply, with the time to its first audio and the price of the turn.
We gave Qwen Audio 3.1 TTS Plus, Qwen Audio 3.1 TTS, Gemini 2.5 Pro TTS and StepAudio 3 TTS the same narration and direction, one shot each. Listen to all four, with the price and time of each request.
A document OCR model and three vision chat models on the same three pages (an invoice, a two-column report, and lecture notes with equations), one shot each, with accuracy, what comes back, and price per page.
API Access
LightOnOCR-3 4B is available on EmpirioLabs: send a PDF or an image and get each page back as Markdown, with tables, math in LaTeX, headers and footers, and the position of every block on the page.
Stable Audio 3 Medium is available on EmpirioLabs: describe music or a sound and get a 44.1 kHz stereo clip from 1 second to more than six minutes, in MP3, WAV, FLAC, or Ogg. Powered by Stability AI.
MOSS-Image Turbo is available on EmpirioLabs: describe an image, put the words you want in quotation marks, and get signs, posters, and labels with the text spelled out, in English or Chinese.
Four image models on three prompts that need lettering, an English trail sign, a jazz poster, and a Chinese shop sign, one shot each at 1024x1024, with prices.
Three audio models on the same prompts, a 30-second lo-fi beat and a 10-second steam train, one shot each, with lengths, formats, and prices side by side.
We gave MiniMax M3.1 Flash, DeepSeek V4.1 Flash 1017, Qwen3.8 Flash and MiMo V2.6 Flash the same four coding prompts, one shot each at the highest reasoning setting, and measured time, output tokens and price, with every result rendered live.
MiniMax M3.1 Flash is available on EmpirioLabs: a fast multimodal model for coding and agents with a 1M token context, adjustable reasoning depth, image and video input, and built-in web search.
We gave Qwen Audio 3.1 TTS Next and StepAudio 3 Gen the same written radio scene, with two speakers, a sound effect and background music, one shot each. Listen to the results, with the price of each request.
Showing 25–36 of 155 articles
Explore our models, or contact us about business inquiries, custom deployments, or anything else.