The Pruna-Qwen-Image-2.1 Space is now live! You can now test our few-step LoRA adapters for Qwen-Image-2.1 by @Alibaba_Qwen directly in your browser, no setup or code required. - 5 steps for maximum speed - 8 steps for the best balance of speed and quality - Text-to-image and image editing support Put it to the test and show us the results. B𝘦𝘵𝘵𝘦𝘳 𝘷𝘦𝘳𝘴𝘪𝘰𝘯𝘴 𝘰𝘯 𝘵𝘩𝘦 𝘸𝘢𝘺. 🚀 Test the Pruna-Qwen-Image-2.1 Space on @huggingface by @_akhaliq: Check Pruna-Qwen-Image-2.1:

Today, we open-source Pruna-Qwen-Image-2.1, a set of a few-step LoRA adapters that make Qwen-Image-2.1 by @Alibaba_Qwen up to 6.3× faster for image generation and editing. - 𝗛𝗼𝘄 𝗲𝗳𝗳𝗶𝗰𝗶𝗲𝗻𝘁 𝗶𝘀 𝗶𝘁? Generate or edit images in just 5 or 8 steps instead of 40. Choose 5 steps for maximum speed or 8 steps for the best balance. - 𝗛𝗼𝘄 𝗴𝗼𝗼𝗱 𝗶𝘀 𝗶𝘁? The 8-step adapter is our recommended default. It supports text-to-image and single- or multi-image editing at 1K resolution, with up to three reference images. - 𝗛𝗼𝘄 𝗱𝗼𝗲𝘀 𝗶𝘁 𝘄𝗼𝗿𝗸?The LoRA adapters load directly on top of Qwen-Image-2.1, while the pipeline remains unchanged. Each adapter uses its own optimized sigma schedule and runs without CFG. 𝘛𝘩𝘪𝘴 𝘪𝘴 𝘫𝘶𝘴𝘵 𝘰𝘶𝘳 𝘧𝘪𝘳𝘴𝘵 𝘳𝘦𝘭𝘦𝘢𝘴𝘦, 𝘸𝘪𝘵𝘩 𝘦𝘷𝘦𝘯 𝘣𝘦𝘵𝘵𝘦𝘳 𝘷𝘦𝘳𝘴𝘪𝘰𝘯𝘴 𝘰𝘯 𝘵𝘩𝘦 𝘸𝘢𝘺. 🚀 Want to try the fast open-source integration? Check it here on Diffusers: Want to try the fastest image generation and editing endpoint? Check P-Image-Ideogram and P-Image-Edit here on API:

The Pruna Build Program is open for teams shipping real AI products. Everyone can try in Stage 0. Then, start with $300 in free credits after signup, and unlock up to $ 3,900 total as your usage grows. No heavy application: add billing info, build, and unlock the next stage when you hit usage milestones. Built for SaaS, API, and teams who care about cost, speed, and efficiency. You can also get early access to our P-Models, visibility, and extra support. Read how it works: Apply in 2 minutes: Applications are reviewed monthly (third week of each month). 🌐 Website: 📚 Documentation: 🧩 API:

No set. No recapture. Same face, same mic. The vocal is already on the track. P-Video-2 generates a whole new video from text, image, or audio-conditioned input. A jump in quality and efficiency for videos up to 20 seconds. Standard from $0.025/s of output video · Draft mode from $0.015/s What's the setup? 👀 - Model: P-Video-2 - Input image: Artist on stage with a microphone - Input audio: Imported music track - Input prompts: A young woman sings the song, natural mouth shapes, believable breathing, and facial expressions that reflect the emotion and intensity of the vocals. She maintains a confident stage presence while making subtle rhythmic movements. 👉 Free Playground: 📚 Documentation: 🧩 API:

P-Video-Edit and P-Video-Edit Draft are in the same performance band on Design Arena. Draft scores 1170 Elo vs 1163 for Standard, while cutting average generation time from 64.1s to 28.5s. That puts P-Video-Edit Draft on the Pareto frontier for preference × speed, as well as price × speed. Benchmark by @DesignArena. 👉 Free Playground: 📚 Documentation: 🧩 API: 🥇 P-Bench:

The product is ready. The creator is not. Do you wait on a shoot, or generate the ad from the still? P-Video-2 generates a whole new video from text, image, or audio-conditioned input. A jump in quality and efficiency for videos up to 20 seconds. Standard from $0.025/s of output video · Draft mode from $0.015/s What's the setup? 👀 - Model: P-Video-2 - Input image: Creator holding a serum bottle in a bathroom - Input prompts: The woman holds the serum bottle toward the camera and speaks enthusiastically: "Okay, this serum actually fixed my dry patches in one week." Subtle head movement, natural blinking, static phone-camera framing, warm bathroom background unchanged. Audio: soft trending lo-fi hip-hop beat playing under her voice, her voice clear and close-mic, very light bathroom room tone — never voice-only. 👉 Free Playground: 📚 Documentation: 🧩 API:

@PrunaAI ships P-Video-2 today, live on Eden AI from day one. Draft mode is the best bit: video is billed per second of output, so the cost was never the clip you ship, it is the ten you deleted first. Fast, low-cost video generation, now available worldwide through Eden AI. ↓

P-Video-2 from @PrunaAI is now live on Runware 🎬 Pruna AI’s latest T2V, I2V, and audio-conditioned video model is here on Runware, delivering up to 20s outputs at 1080p, 48fps. Draft mode at half the per-second rate for fast iteration. From $0.025/s full · $0.015/s draft

New skins, new scenarios, same deadline. No time to record the gameplay again. P-Video-Edit edits the recording you already have. One source video (up to 15s), one instruction. Draft mode from $0.025/s of output video · Standard from $0.045/s. What's the setup? 👀 - Model: P-Video-Edit - Input video: Orange poly runner on a grey city rooftop course - Input prompts: 1) Change only the main player character from orange to vivid electric blue, persistently and consistently in every frame from beginning to end. 2) Change the entire game environment into a colorful candy world, persistently from the first frame to the last. Replace the gray city with candy towers, lollipops, gumdrops, and cotton-candy clouds. 👉 Free Playground: 📚 Documentation: 🧩 API:

How does P-Video-Edit performs on quality, speed, and price? It’s Pareto optimal across all three. 👀 We benchmarked quality, speed, and price on P-Bench: • P-Video-Edit Standard: ~1000 Elo · $0.045/s · 23.18s gen time / output second • P-Video-Edit Draft: ~994 Elo · $0.025/s · 11.48s • Grok Imagine Video: ~974 Elo · $0.05/s · ~11.0s • Gemini Omni Flash Edit: ~1054 Elo · $0.13/s · ~13.8s • MiniMax H3: ~1060 Elo · $0.06/s · ~58s Draft mode gets within 6 Elo of Standard mode while costing 44% less and generating ~2× faster. 🥇 Full benchmark:

You locked a beauty creator take, then the shade lineup changes and marketing wants a red variant the same day. P-Video-Edit edits the footage you already have. One source video (up to 15s), one instruction. Draft mode from $0.025/s of output video · standard from $0.045/s. What's the setup? 👀 - Model: P-Video-Edit - Input video: Beauty creator with gloss bottle and lip shot - Input prompt: Change only two color elements: turn the gloss liquid inside the bottle from pale rose pink to rich red throughout the video, and make the gloss on the person's lips red in the final lip shot. 👉 Free Playground: 📚 Documentation: 🧩 API:

🎓 Using P-Video-Edit? We've just dropped five new creator guides to boost your outputs! Editing video: Removing objects from video: Editing with reference images: Draft mode and reproducible reruns: Prompt upsampling:

p-video-edit by @PrunaAI is now available on Design Arena! p-video-edit enables instruction-based video-to-video editing with support for up to 4 reference images. Edit videos up to 15 seconds with simple prompts, swap styles and objects, and preview your edits faster with Draft mode. Results to follow shortly! Congrats to the @PrunaAI team on the launch!

P-Video-Edit @PrunaAI is now live on WaveSpeed. Instruction-based video editing, powered by your prompt and references. → Edit videos up to 15s → Add up to 4 image references → Swap objects, styles & more One source video. One prompt. A whole new version.

Happy to announce P-Video-Edit: instruction-based video-to-video editing with up to 4 references for just $0.025/s. • 𝗜𝗻𝗽𝘂𝘁: One source video (up to 15s) + one edit prompt • 𝗥𝗲𝗳𝗲𝗿𝗲𝗻𝗰𝗲𝘀: Up to 4 optional images for supported style and object swaps • 𝗣𝗿𝗶𝗰𝗶𝗻𝗴: Draft from $0.025/s · standard from $0.045/s • 𝗗𝗿𝗮𝗳𝘁 𝗺𝗼𝗱𝗲: Faster preview before you commit to the full render Don't reshoot it. Just edit. Available via our inference partners @scenario_gg, @Replicate, @WiroAI, @wavespeed_ai, @prodialabs, @Runware, @togethercompute , @runpod, @Cloudflare, @eachlabs, @TellersAI, @Segmind_ai Validated by our benchmark partners @DesignArena, @datapointai, @Rapidata 👉 Free Playground: 📚 Model page: 🧩 API:

P-Image-Ideogram, built with @ideogram_ai, now has a Very High mode. Very High uses test-time scaling to push final image quality further. Built for hero assets, difficult prompts, and quality-critical generations when people and products have to pop. • ~5.55s at 1K · $0.033/image (1K) · $0.066/image (2K) • Five modes from Very Low ($0.003) through Very High Again, benchmarks agree: This new very high mode is on efficiency-quality based on human and automated evals from @datapointai, @DesignArena, and P-Judger. Full benchmarks results follow in a later post on P-Bench. Available on @cloudflare @ComfyUI @GammaApp @inference_sh @kittldesign @LeonardoAi @lovart_ai @magnific @picsart @replicate @runware @scenario_gg @togethercompute @wavespeed_ai @wiroai 👉 Free Playground: 🧩 API: 📚 Model page: