StepFun
@StepFun_ai
Followers
2K
Following
59
Media
18
Statuses
70
Scale-up possibilities for everyone. Discord: https://t.co/qVxsj6WZON
Joined February 2025
🎙️ “Switch voices like flipping a script?” Now that’s your flex. Meet Voice Rag in Step-Audio 2 — where roleplay goes from 0 to 100 in a single command. 👉 Say “Become a CEO” — the voice morphs instantly. No lag, no awkwardness — just next-level character-switching magic.
3
3
32
RT @vllm_project: This amazing Attention-FFN disaggregation implementation from @StepFun_ai , achieves decoding throughput of up to 4,039 t….
0
50
0
Step 3 will be open-sourced on July 31st! Coming Soon🤟.
The wait is over!. Meet Step 3 — the groundbreaking multimodal LLM from StepFun!. 🚀 MoE architecture (321B total params, 38B active).💡 Rivals OpenAI o3, Gemini 2.5 Pro, and Claude Opus 4 in performance.🖥️ Optimized for China’s domestic AI chips. StepFun just announced: Step 3
8
14
128
🚀 StepFun Deep Research achieves SOTA with a 70% pass rate on xBench-DeepSearch and excels on BrowseComp! 🏆.An end-to-end Multi-Agent system for automated, in-depth research. 📄 Generate Deep Research Reports: From hundreds of sources to a full report. ✅ Intelligent
5
35
173
🚀 Last but not least, we also released two new benchmarks!✍️. 🔗 StepEval-Audio-Paralinguistic —— Assessing paralinguistic information understanding.🔗 StepEval-Audio-ToolCall —— Evaluating tool-calling capabilities).
huggingface.co
0
0
9
💡 Breakthrough Capabilities & Technical Highlights of Step-Audio 2 ✨✨. ✅ SOTA ASR: English, Mandarin and Chinese dialects.✅ SOTA audio understanding: sound, music & non-verbal cues.✅ Paralinguistic information reasoning: analyzing emotions before responses.✅ Multimodal
1
0
6
🏆 Outperforming GPT-4o: Meet Step-Audio 2!.An end-to-end LALM designed for industry-strength audio understanding and speech interaction. ✨ Emotion-aware reasoning.✨ Switch timbres with natural language.✨ Intelligent speech conversation. 🚀 SOTA on LibriSpeech, MMAU,
7
19
120
RT @yanawei_: 🔥 Thrilled to release our new multimodal RL work: Open Vision Reasoner!.A powerful 7B model with SOTA performance on language….
0
6
0
RT @CyouSakura: We are excited to introduce Open Vision Reasoner (OVR) 🚀 — transferring linguistic cognitive behavior to unlock advanced vi….
0
28
0
Paper here:
huggingface.co
Open Vision Reasoner. Transferring Linguistic Cognitive Behavior for Visual Reasoning
0
2
15
RT @rtedevcommunity: #Voice_Agent 模型分享：@StepFun_ai 开源了其最新的大型音频语言模型 Step-Audio-AQAA，并已在 Hugging Face 上线。.Step-Audio-AQAA 是一个完全端到端的模型，专注于音频问答….
0
2
0
RT @Xianbao_QIAN: A new large audio language model from @StepFun_ai is now open sourced on @huggingface . - Based on LLM 130B Step-Omni.- M….
0
27
0
🔥🔥Strongly recommended this youtube tutorial of NEW released #ACEStep open-source AI music generator with @ACEStudio_en !! .Very detailed and Newbie-friendly. ⚡️⚡️Thanks for @aisearchio !!.👉👉 Check it out:
0
1
10
ComfyUI 🔥🔥🔥 Thx for supporting in @ComfyUI . Really recommended to try the workflow. Looking forward to see amazing works from community !!!.
Stable Diffusion Moment of Audio?? Ace-Step Audio Model just dropped and is now natively supported in ComfyUI. ACE-Step is an open-source music generation model jointly developed by @ACEStudio_en and @StepFun_ai.
0
0
7
🥳🥳🥳Try it at demo page:
ace-step.github.io
Just found SOTA open-source music generation model!.Songs、Instrumentals、Multiple Languages、Rexmix、Repaint、Edit. And they Release training code, Release LoRA training code 🔥
1
1
22
RT @ACEStudio_en: We’re excited to release ACE-Step / ACE-Step-v1-3.5B, a fast, versatile DiT-based foundation model for music generation t….
0
56
0
RT @zsakib_: i've just added @StepFun_ai's Step1X-Edit onto @replicate (it's been trending on hf spaces).> it lets you edit images with tex….
0
3
0
