Fast AI inference. 200+ models via one API. OpenAI/Anthropic compatible. 💬discord.gg/X5Z29RvKkD
Singapore
- Tweets 551
- Following 278
- Followers 2,777
- Likes 1,015
Pinned Tweet
🌊 Clear Your GLM 5.2 Spend. Up to $1,000 Voucher
🍺 SiliconFlow Summer Rush-GLM 5.2 Week is LIVE
From 20:30:00 on June 29 to 20:30:00 on July 6, PDT
How it works ↓
🎟️ Entry: run GLM 5.2 on SiliconFlow + post use-case on X + fill the register form
🏆 Then climb: the more GLM 5.2 you run, the higher you rank
Top 1 Builder gets:
👑 Your GLM 5.2 spend this week, refunded by voucher — up to $1,000
💰 Extra $50 voucher
📢 Feature your work on SiliconFlow's X with personalized winner poster
👑 The exclusive GLM 5.2 Token Legend Discord title
Plus:
⚡ Early Bird Prize: post early for extra voucher
🎲 Lucky Draw: every valid entry has a chance to win
👇 More details? Full guide in the thread
5
2
12
11,709
硅基流动已上线Qwen2.5加速版，欢迎大家体验🥳
🚀Now it is the time, Nov. 11 10:24! The perfect time for our best coder model ever! Qwen2.5-Coder-32B-Instruct!
Wait wait... it's more than a big coder! It is a family of coder models! Besides the 32B coder, we have coders of 0.5B / 1.5B / 3B / 7B / 14B! As usual, we not only share base and instruct models, we also provide quantized models in the format of GPTQ, AWQ, as well as the popular GGUF! 💖
👉🏻Blog: qwenlm.github.io/blog/qwen2.…
👉🏻Tech Report: arxiv.org/abs/2409.12186
👉🏻Hugging Face: huggingface.co/collections/Q…
👉🏻ModelScope: modelscope.cn/collections/Qw…
👉🏻Kaggle: kaggle.com/models/qwen-lm/qw…
👉🏻GitHub: github.com/QwenLM/Qwen2.5-Co…
👉🏻Demo [chat]: huggingface.co/spaces/Qwen/Q…
👉🏻 Demo [Artifacts]: huggingface.co/spaces/Qwen/Q…
The flagship model, Qwen2.5-Coder-32B-Instruct, reaches top-tier performance, highly competitive (or even surpassing) proprietary models like GPT-4o, in a series of benchmark evaluation, including HumanEval, MBPP, LiveCodeBench, BigCodeBench, McEval, Aider, etc. It reaches 92.7 in HumanEval, 90.2 in MBPP, 31.4 in LiveCodeBench, 73.7 in Aider, 85.1 in Spider, and 68.9 in CodeArena!
5
11
111
54,502
🥳MiniMax-M1-80k (456B) is now available on SiliconCloud, supporting up to 128K context length!
The pricing is set at ¥4/M Tokens for input and ¥16/M Tokens for output, with new users receiving ¥14 in credits for free usage.
✨ Key Model Highlights:
1. Up to 1M token input support
2. Highly efficient inference: Generating 100K tokens with just 25% of DeepSeek R1's FLOP cost
3. MoE architecture + Lightning Attention, ideal for agents, code generation, and complex reasoning
👀Explore it here: cloud.siliconflow.cn/models
📃API Docs: docs.siliconflow.cn/cn/api-r…
👨🔧Integration Guide: docs.siliconflow.cn/cn/userc…
🦖Excited to collaborate with the MiniMax team @minimax_ai and share this journey together — looking forward to seeing how this model revolutionizes AI-powered applications!😜
#minimax #SiliconCloud #AImodels #OpenSource #AItool #GenerativeAI #AGI
19
2
19
6,963
当天就有教程了！！感谢 orange 老师！！
今天硅基流动的 API 总算是上线了，这是目前为数不多的稳定、高速、满血版的 DeepSeek R1 API。
但是 R1 很特别，如果不配置好的话，R1 的效果会大打折扣，甚至会直接跳过思考过程。
所以写篇教程跟大家分享心得，避免大家踩坑。
《DeepSeek R1 API 获取和使用指南》
mp.weixin.qq.com/s/u_ODtvzhv…
5
4
22
5,243
Join the CAMEL AI open-source community and explore AI together!
🚀 Exciting News! 🚀
CAMEL AI has officially integrated with @SiliconFlowAI , a cutting-edge AI Infra platform designed for scalable, standardized, and high-performance AI model services! 🔥
With this integration, users can now seamlessly access @deepseek_ai R1 via SiliconFlow, enabling fast and cost-efficient model fine-tuning and deployment. Whether you're a developer or an enterprise, this one-stop platform empowers you to focus on innovation while we handle the AI infra for you.
Try it out today and supercharge your AI workflows! ⚡
👉 Explore more here: github.com/camel-ai/camel/pu…
1
11
1,426
【LangChain x SiliconFlow】meetup in Beijing was a great success! 🥳
Over a few hours, we enjoyed amazing tech talks and lively interactions—everyone was so enthusiastic!
A big thank you to Harry @zhanghaili0610 and the Langchain team! @langchain
1
14
3,803
😉Qwen2.5 models are now live on Siliconflow!! 😉(Accelerated inference version.) 🦾🦾🦾
Playground：
•cloud.siliconflow.cn/s/Qwen2… •cloud.siliconflow.cn/s/Qwen2… •cloud.siliconflow.cn/s/Qwen2… •cloud.siliconflow.cn/s/Qwen2…
API：
docs.siliconflow.cn/referenc…… pic.x.com/zegf0xwvud
1
1
12
1,400
🤗Qwen-32B-Preview is now live on SiliconCloud🤗
Try it online: cloud.siliconflow.cn/playgro…
See API Doc: docs.siliconflow.cn/api-refe…
🪶We're releasing a preview of QwQ /kwju:/ — an open model designed to advance AI reasoning capabilities.
Blog: qwenlm.github.io/blog/qwq-32… Model: hf.co/Qwen/QwQ-32B-Preview
Demo: hf.co/spaces/Qwen/QwQ-32B-pr…
QwQ has preliminarily demonstrated remarkable capabilities, especially in solving some challenges in mathematics and coding. As a preview release, we acknowledge its limitations. We earnestly invite the open research community to collaborate with us to explore the boundaries of the unknown!
4
12
8,483
We’ve Completed Our Series A Funding — Tens of Millions of USD Raised!
We’re excited to announce that we’ve successfully completed our Series A funding round, raising tens of millions of USD. This round was led by Alibaba Cloud, with continued strong support from Sinovation Ventures.
“As a team dedicated to advancing AI infrastructure, we’ve continuously driven progress through technical breakthroughs and product innovation,” said our founder, Dr. Yuan Jinhui. “With the rapid growth of open-source models like Qwen and DeepSeek — and soaring demand for AI inference — we’ve entered a phase of accelerated expansion. With this new funding, we’ll accelerate R&D and expand our services to more developers and enterprises worldwide, striving to become the go-to platform for generative AI development.”
We’re focused on solving one of the core challenges in generative AI: the high cost and complexity of large-scale inference. Our self-developed high-performance inference engine significantly improves computational efficiency and is deeply optimized for both international and emerging hardware platforms.
In early 2025, we launched DeepSeek-R1 and V3, inference services based on alternative compute infrastructure. These services match the performance and cost-efficiency of mainstream GPU deployments, proving that large models can be commercially viable on alternative hardware.
To meet dynamic demand in real-world AI workloads, we’ve developed a one-stop heterogeneous compute orchestration platform that enables elastic scheduling and intelligent scaling. This platform helps unify fragmented compute resources and transforms infrastructure from a constraint into a productivity driver.
For developers, our SiliconCloud platform provides access to over 100 open-source foundation models. It provides end-to-end support — from fine-tuning and hosting to deployment — empowering teams to build AI products with ease.
Over the past year, SiliconCloud has rapidly become one of the most developer-friendly generative AI platforms:
1. 6 M+ total users
2. Thousands of enterprise clients
3. Over 100 billion tokens generated daily
We’ve also launched BizyAir, a creative AI workflow platform that integrates cloud GPUs with local tools like ComfyUI. With rich templates, custom model support (including LoRA), and seamless workflow design, it’s already being used in advanced AI video and image generation pipelines.
Our product suite now includes API services, dedicated instances, software subscriptions, and AI appliances — powering applications across language models, image synthesis, video generation, and more. We proudly serve a growing number of enterprise customers across sectors including technology, finance, manufacturing, and design.
We’re just getting started. In the months ahead, we’ll continue pushing the boundaries of AI infrastructure — making it more affordable, accessible, and globally deployable.
3
7
689
DeepSeek Janus-Pro-7B model is now live on SiliconCloud! 🤗
🚀 Introducing Janus-Pro-7B:
1. Vision decoupled, independent channels.
2. Simple, flexible, efficient.
3. Outstanding performance.
Try it on: cloud.siliconflow.cn/models
#JanusPro #MultimodalLearning #AI
1
1
10
1,987
📢The accelerated version of Fish-speech-1.5 is now on SiliconCloud!!🐟💨
Dive in and experience lightning-fast speech synthesis like never before🚀: cloud.siliconflow.cn/models?…
Introducing Fish Speech 1.5 🎉 - Making state-of-the-art TTS accessible to everyone!
Highlights:
- #2 ranked on TTS-Arena (as "Anonymous Sparkle")
- 1M hours of multilingual training data
- 13 languages supported, including English, Chinese, Japanese & more
- <150ms latency with high-quality instant voice cloning
- Pretrained model now open source
- Cost-effective self-hosting or cloud options
Let's check out the details 🧵⬇️
11
860
🌑Kimi-K2-Instruct is now live on SiliconFlow!
🌖Open-source MoE model with 1T total and 32B active parameters.
🌖Strong at coding and Agent tasks, with solid benchmark results in programming, tool use, and reasoning.
🌖Supports up to 128K context length.
💰Pricing: $0.58 /M Tokens (input), $2.29 /M Tokens (output)
👏New users get $1 free credit!
👉 Try online: siliconflow.com/models/moons…
👉 Developer API docs: docs.siliconflow.com/en/user…
🚀 Hello, Kimi K2! Open-Source Agentic Model!
🔹 1T total / 32B active MoE model
🔹 SOTA on SWE Bench Verified, Tau2 & AceBench among open models
🔹Strong in coding and agentic tasks
🐤 Multimodal & thought-mode not supported for now
With Kimi K2, advanced agentic intelligence is more open and accessible than ever. We can't wait to see what you build!
🔌 API is here: platform.moonshot.ai
- $0.15 / million input tokens (cache hit)
- $0.60 / million input tokens (cache miss)
- $2.50 / million output tokens
🔗 Tech blog: moonshotai.github.io/Kimi-K2…
🔗 Weights & code: huggingface.co/moonshotai
🔗 Github: github.com/MoonshotAI/Kimi-K…
Try it now at Kimi.ai or via API!
2
8
1,635
硅基花样流动ComfyUI & BizyAir征稿活动的最高奖巅峰艺术奖作者LogicAI带着他的作品来啦!!!👍👍
作品同时被我们工程师带到了东京CCS大会进行了展出，收获了大家的一致好评!!!🤗🤗
除此之外，LogicAI创作的使用视频质量也是超一流!!不愧是动画制作出生的大佬!!!欢迎大家一起欣赏👏👏
诚实卡片生成器（Honest Deck Maker）是一个能够全自动生成具有讽刺意味的游戏卡片的ComfyUI工作流，支持多语言输入输出，目前100%适配日文和中文。用户只需要输入卡片名称和选定语言，剩下部分就能全自动执行。项目灵感来源于 Screen Junkies 的《诚实预告片》(Honest Trailers)系列，并且结合了《炉石传说》的视觉设计美学与李继刚先生编写的《汉语新解》提示词中的语言哲学。自定节点方面，本工作流采用了 SIliconflow 提供强大的 LLM 能力并整合 BizyAir 提供的 FLUX 文生图云端推理能力，实现了在纯CPU环境下也能快速生成图片，欢迎下载工作流来玩：openart.ai/workflows/tuatara…
1
10
1,954
Two more new models are here too.👐
Qwen2.5-Coder-7B: cloud.siliconflow.cn/s/Qwen2…
Qwen2.5-Math-72B: cloud.siliconflow.cn/s/Qwen2…
😉Qwen2.5 models are now live on Siliconflow!! 😉(Accelerated inference version.) 🦾🦾🦾
Playground：
•cloud.siliconflow.cn/s/Qwen2… •cloud.siliconflow.cn/s/Qwen2… •cloud.siliconflow.cn/s/Qwen2… •cloud.siliconflow.cn/s/Qwen2…
API：
docs.siliconflow.cn/referenc…… pic.x.com/zegf0xwvud
1
1
10
848
🥳Qwen3-235B-A22B-Instruct-2507 is now live on SiliconFlow.
With the model, you can expect:
1️⃣High-Speed Inference: Optimized for lower latency and higher throughput.
2️⃣Cost-Effective Pricing: $0.35/M tokens (input) and $1.42/M tokens (output).
✨ Key Model Highlights:
1.Enhanced General Capabilities: Smarter reasoning, math, coding, and tool use.
2.Better User Alignment: More helpful and intent-aligned responses.
3.Expanded Multilingual Knowledge: Broader coverage across languages and domains.
4.Extended Context Understanding: 256K long-context understanding capabilities.
👏New users get $1 free credit!
👉 Explore here: siliconflow.com/models/qwen-…
@Alibaba_Qwen
1
9
630
欢迎大家免费报名参加🥳🥳🥳
位置在北京海淀搜狐网络大厦，丰富的活动内容和奖品等着大家，咱们10月26日不见不散!!!🧡（其他具体信息详见图片）
🇨🇳 LangChain Ambassador @zhanghaili0610 is teaming up with SiliconFlow to host an in-person LangChain meetup in Beijing!
Come join us to dive into LangChain, learn about AI infrastructure, and connect with fellow developers.
Register here using the QR Code ➡ lu.ma/mo5nf9wx
2
8
2,572
🚀 BizyAir Now Supports FLUX.1 Tools for Enhanced Image Processing and Control!
🌟 Upgrade to the latest version of BizyAir and experience the power of FLUX.1 Tools firsthand!
🔥 github.com/siliconflow/BizyA…
🔑 bizyair.siliconflow.cn/
#BizyAir #FLUX1Tools #FLUX #ImageEditing #AI
9
520
🎉GLM-4.5 and GLM-4.5-Air are now live on SiliconFlow!
With SiliconFlow's GLM-4.5 API, you can expect:
1️⃣Cost-Effective Pricing:
GLM-4.5 at $0.5/M input tokens and $2/M output tokens;
GLM-4.5-Air at $0.14/M input tokens and $0.86/M output tokens.
2️⃣Extended Context Window: 128K for complex tasks.
✨Key Capabilities：
1.SOTA Performance: Leading in reasoning, coding, and agentic tasks.
2.MoE Architecture: GLM-4.5 (355B/32B), GLM-4.5-Air (106B/12B), optimized for efficiency.
3.Hybrid Inference: Complex tasks & instant responses.
👏New users get $1 free credit!
👉 Explore here:
GLM4.5: siliconflow.com/models/zai-o…
GLM4.5-air: siliconflow.com/models/zai-o…
Developer API docs: docs.siliconflow.com/en/user…
Congrates！ @Zai_org
1
9
608
🥳SiliconFlow x DeepSeek🥳: Accelerated Version of DeepSeek-VL2 Launched First on SiliconCloud now!!
🎉 DeepSeek-VL2 is here! Our next-gen vision-language model enters the MoE era.
🤖 DeepSeek-MoE arch + dynamic image tilling
⚡ 3B/16B/27B sizes for flexible use
🏆 Outstanding performance across all benchmarks
🧵 1/n
9
690
🚀 Qwen3-235B-A22B-Thinking-2507 is now available on SiliconFlow！
With SiliconFlow's API, you can expect:
1️⃣Cost-Effective Pricing: $0.35/M tokens (input) and $1.42/M tokens (output).
2️⃣Extended Context Window: 256K context window for complex tasks.
✨Key Capabilities：
1.SOTA Reasoning Performance: Stronger logic, math, coding, and academic task performance.
2.Enhanced General Capabilities: Improved instruction following, tool use, and response alignment.
🎁 $1 in free credits for all new users!
Get Started Immediately ！
Explore: siliconflow.com/models/qwen-…
Developer API docs: docs.siliconflow.com/en/user…
1
5
475
🚀QVQ-72B-Preview Turbo version is now live on SiliconFlow!
🤟QVQ excels in NLP and multimodal benchmarks (e.g., MMLU, VQAv2), surpassing top models like GPT-4.
Try it now at: cloud.siliconflow.cn/models?…
1
7
606
The first 1k stars are reached. This a milestone for the onediff team! Thanks for the great feedback from the GenAI community: github.com/siliconflow/onedi…
1
1
7
1,133
🚀 New Episode Alert! 🎥
BizyAir Ep.7 is LIVE! Dive into the world of InstantID & IPAdapter as we explore their unique applications and showcase real-world scenarios.
👉🏻 bilibili.com/video/BV1qTmRYC…
🥳 Visit our website: bizyair.siliconflow.cn/index…
#AI #BizyAir #TechTutorials
6
431
太牛啦！！！
🚀 DeepSeek-R1 is here!
⚡ Performance on par with OpenAI-o1
📖 Fully open-source model & technical report
🏆 MIT licensed: Distill & commercialize freely!
🌐 Website & API are live now! Try DeepThink at chat.deepseek.com today!
🐋 1/n
4
6
1,249
🎙️CosyVoice 2 is now live on SiliconCloud!🎙️
150ms real-time speech synthesis with support for mixed languages and dialects.🏂
Give it a spin and hear the difference at: cloud.siliconflow.cn/playgro…
5
582
🎉 New Episode Alert!
🔥Dive into Episode 4 of our BizyAir tutorial series! 🔥
This episode introduces how to set up an image-to-image framework, and wraps up with a guide on using JOY caption's backtracking node.
👉 bilibili.com/video/BV1WrmVYu…
#ComfyUI #StableDiffusion #FLUX
5
338
不是吧不是吧🫢还有谁不知道我们上线了FLUX🥺
FLUX 是新出的目前最强开源图像模型
这几天在找哪个地方可以试用
找了一圈发现硅基流动的体验中心就可以用。
而且还是限时免费的随便用。。。
这羊毛不薅岂不是可惜，来，用起来
放一张 MJ 和 FLUX 的对比图，光线不如 MJ 梦幻，偏写实一些
cloud.siliconflow.cn?referre…
2
2
6
997
Be one of the first to try the new LTX-Video on SiliconCloud!
⚡ This free, open-source model lets you create amazing content in seconds. Start to generate videos effortlessly with this powerful model!!
Free for a limited time: cloud.siliconflow.cn/playgro…
#AI #VideoEditing
3
517
🎉 Episode 5 of BizyAir Tutorial Series is now available!
This video introduces how to use the inpainting function to repair details in the generated image, in order to improve the general quality of images.
👉🏻 bilibili.com/video/BV1nqyRYG…
#ComfyUI #StableDiffusion #FLUX
5
344
🚀 New BizyAir Tutorial Alert!
In Episode 6, we dive into ControlNet models to enhance control over image movement, depth, and detail, enabling large language models to create visuals that meet precise specifications.
👉🏻 bilibili.com/video/BV1x7DdYi…
#ComfyUI #FLUX #StableDiffusion
5
335
📢 New on SiliconFlow: Hunyuan-A13B-Instruct is live!
1️⃣MoE: 80B total, 13B active
2️⃣Code, reasoning, agent tasks, long context
3️⃣$0.14/M tokens (in), $0.57/M tokens (out)
🔍 Get $1 credit and try it now: siliconflow.com/models/tence…
#AI #LLM #SiliconFlow #AIModels
5
517
📢 New on SiliconFlow: Baidu ERNIE-4.5-300B-A47B is live! ✨
1️⃣ 300B-class Chinese-centric models
2️⃣ Code, reasoning, long context, bilingual chat
3️⃣ $0.14/M tokens (in), $0.57/M tokens (out)
⚡️ Standard OpenAI-style APIs, easy integration.
🔗 Try it now: siliconflow.com/models/baidu
5
534
🎬🚀✨ New Model Drop: Wan 2.2 Series just landed on SiliconFlow!
This upgrade to Wan’s visual generative models means stable, realistic, cinematic videos are now within reach:
🌟 Cost-Effective Pricing – $0.29 / video (T2V & I2V) 🌟 Resolution Support – 480P & 720P (5s)
🌟 T2V-A14B – High-quality video gen, top on Wan-Bench 2.0
🌟 I2V-A14B – Stable, coherent motion + diverse stylized scenes
🌟 Beats closed-source models like Sora & Hailuo 02 across key benchmarks
Key innovations include:
🔑 MoE architecture – Efficient diffusion with expert separation
🎬 Cinematic aesthetics – Lighting, composition, tone finely controllable
⚡ Complex motion gen – +65.6% more images, +83.2% more videos vs. Wan 2.1
Perfect for 🎨 concept art, 📺 commercial visuals, 🎥 cinematic content creation, and next-gen creative AI 🚀✨
👉 Try it now: cloud.siliconflow.com/models 👉 API Docs: docs.siliconflow.com/cn/api-…
2
5
518
🚀 Release of HunyuanVideo: The King of Open-Source Video Generation Models 🎥
🌟 We’re thrilled to announce the launch of the inference-accelerated version of HunyuanVideo, an incredible breakthrough in video generation!
👉🏻 cloud.siliconflow.cn/playgro…
4
333
Level up your AI art game with SD3.5 Large ControlNet Tools on BizyAir!
🎨 Blur, Canny, and Depth models are now available for easy cloud-based access. Precise control over your AI images with their incredible performance！
👉🏻 github.com/siliconflow/BizyA…
#StableDiffusion #BizyAir
4
464
We are glad to share that #OneDiff SDXL inference has been integrated into fal.ai playground! SDXL inference at the speed of thought. Let's Check it out!
fal.ai/models/onediff-sdxl @fal_ai_data #SDXL #TextToImage #GenerativeAI
4
292
FLUX.1 Kontext Dev is now live on SiliconFlow! 🚀
With ultra-fast, context-aware capabilities, it's revolutionizing real-time image editing. 🔥
At just $0.015 per image, try the most cost-effective solution on the market.
Explore here: siliconflow.com/models/black…
Big thanks to the @bfl_ml team for their amazing contributions to the open-source community!
1
5
410
A new cute kaomoji model from Qwen is out now! 😂🥳👏
We're also hustling to get its turbo version ready for launch!🦾💪
🎄Happy holidays and we wish you enjoy this year. Before moving to 2025, Qwen has the last gift for you, which is QVQ!
🎉 This may be the first open-weight model for visual reasoning. It is called QVQ, where V stands for vision. It just reads an image and an instruction, starts thinking, reflects while it should, keeps reasoning, and finally it generates its prediction with confidence! However, it is still experimental and this preview version still suffers from a number of limitations (mentioned in our blog), which you should pay attention to while using the model. Feel free to refer to the following links for more information:
* Blog: qwenlm.github.io/blog/qvq-72…
* HF: huggingface.co/collections/Q…
* ModelScope: modelscope.cn/models/Qwen/QV…
* Kaggle: kaggle.com/models/qwen-lm/qv…
🚀 It achieves impressive performance in benchmark evaluation, e.g., MMMU, MathVista, etc. But what is more interesting is that it is exciting to see the AI model behaves differently by thinking deeply and reasoning step by step instead of directly providing answers. Yet, it is still a model for preview. It is unstable, it might fall into repetition, it sometimes doesn't follow instruction, etc. We invite you to try the new interesting model and enjoy playing with it! Feel free to shoot us feedback!
4
831
🚀 GLM-4.5V Model Now Live on SiliconFlow
GLM-4.5V, the world’s leading open-source 100B-scale vision reasoning model, is now live on SiliconFlow, built on GLM-4.5-Air, focusing on complex problem solving and multimodal reasoning.
✅With SiliconFlow's GLM-4.5V API, you can expect:
1.Cost-Effective Pricing: GLM-4.5V $0.14/M tokens (input) and $0.86/M tokens (output).
2.Context Length: 65K-token multimodal context window.
3.Native support: Tool Use and Image Input.
📊 Key Highlights:
1.Image Reasoning: Scene understanding, complex multi-image analysis, spatial recognition.
2.Video Understanding: Long video segmentation and event recognition.
3.GUI Tasks: Screen reading, icon recognition, desktop operation assistance.
4.Complex Chart & Long Document Parsing: Research report analysis, information extraction.
5.Grounding: Precise visual element localization.
⚡ Get Started Immediately
🔗 Model Hub: cloud.siliconflow.com/models
📚 API Docs: docs.siliconflow.com/cn/api-…
5
436
‼️Release of Accelerated Video Model Mochi-1-Preview on SiliconFlow‼️
🌟Developed by GenmoAI, this preview version features high-fidelity motion rendering and exceptional prompt adherence, enabling 480p video generation with remarkable quality.
🔗 cloud.siliconflow.cn/playgro…
#AI
2
4
412
So happy to unlock a new scene with 302.AI ! Welcome to try using our API key on it.🥳🤗👏
🚀We are pleased to announce our collaboration with
@SiliconFlowAI
Users can procure or use models of in #siliconflow directly on the 302.AI platform at the same prices as the official SiliconFlow.
SiliconFlow provides the underlying computational support for 302AI, and 302AI provides the upper layer of application capabilities for SiliconFlow. We hope to bring more benefits and value to users on the road to popularising #AI.
Check the detail👉 medium.com/@302.AI/302-ai-co…
#Partnership #collaboration
3
571
Nice try!!! 👏 Welcome to use Milvus to build a RAG system with SiliconFlow!!! 👨🔧👩🔧🩶
🚀 Excited to announce a tutorial on building a RAG pipeline with Milvus and @SiliconFlowAI 🌐
🔍 SiliconFlow's scalable AI infrastructure integrates with Milvus' vector database to enhance your data processing capabilities.
🧠 Learn how to deploy LLMs and embedding models effortlessly with SiliconCloud's MaaS platform.
📚 Use our step-by-step guide and start building your RAG pipeline today! 👉 milvus.io/docs/build_RAG_wit…
#Milvus #SiliconFlow #AI #RAG
2
404
🚀 DeepSeek-V2.5-1210 is here! The most refined V2 model boosts math, coding, writing, & role-playing skills.
Now live on SiliconCloud with accelerated inference—no deployment needed, just call the API!
🔗 cloud.siliconflow.cn/playgro…
#DeepSeek #AI #LLM
3
458
🎨Introducing our new product, BizyAir! 🎨
This lightweight ComfyUI plugin features multiple cloud nodes and diverse workflows, making your tasks more efficient! 🎉
🐌Detailed Tutorial: [session 1]bilibili.com/video/BV1UrxseG…
#BizyAir #ComfyUI #NewProduct
1
3
425
Hello world! We are the #SiliconFlow team dedicated to building #AI infra. We currently focus on high-performance and cost-saving large-scale model inference and serving solutions. More updates from us coming on X. Stay tuned:-) #LLM #GenerativeAI #TextToImage
1
3
483
Our #TextToImage inference engine #OneDiff has just updated to a new version, which allows you to change the image size for most models without recompilation. Additionally, OneDiff ComfyUI nodes are integrated into the ComfyUI-Manager. #AIGC #GenerativeAI teddit.net/r/StableDiffusion…
3
457
#OneDiff v0.12.1 is now released! This update includes:
SOTA performance update for #SDXL and #SVD, fully support dynamic resolution run, fast #LoRA loading and switching for HF #diffusers, accelerate #InstantID and #SDXLlightning, etc. Get the details:
teddit.net/r/StableDiffusion…
3
449
🎊 Release of the turbocharged Llama-3.3-70B-Instruct!
⚡️🔥 Meta's latest Llama 3.3 (70B) is smaller, faster, and just as smart. It's a major leap forward in reasoning, math, and general knowledge at a fraction of the cost.
👉🏻 cloud.siliconflow.cn/playgro…
#AI #LLM #Llama3
3
293
QwQ-32b-Preview's powerful inference capabilities are redefining what's possible. Don't miss out—experience the accelerated version on the SiliconCloud platform!!💪
Try it now: cloud.siliconflow.cn/models?…
3
654
Hey! Appreciate the shoutout.
We’ve got a free tier so you can test stuff out for your 🎹 project. No big red flags to share!
Happy to help if you have any questions—here’s the docs: docs.siliconflow.com/en/user…
3
164
What a great guide! Felix's report provided very comprehensive and clear analyses on optimizing #SDXL. Glad to see that #OneDiff performed well in it and received a high recommendation from him. #indiehackers #GenerativeAI
1
3
426
🚀 Step3 is now on SiliconFlow — The New Standard in Open-source Multimodal Reasoning
With SiliconFlow's Step3 API, you can expect:
1️⃣Cost-Effective Pricing: Step3 $0.57/M tokens (input) and $1.42/M tokens (output).
2️⃣Context Length: Supports 64K context length.
✨Key Capabilities：
1.VLM benchmarks: MMMU 74.2 (↑Gemini 2.5 Flash), Hallusion Bench 64.2 (↑Claude Opus 4)
2.LLM benchmarks: AIME25 82.9, GPQA-Diamond 73.0, LiveCodeBench 67.1
3.Architecture: MFA attention, lower KV cache, faster inference
4.Multimodal encoder: 5B vision model, 1/16 token downsampling
5.System design: AFD pipeline, higher throughput
👏New users get $1 free credit!
🚀 Get Started Immediately
Explore: cloud.siliconflow.com/models
Integrate:docs.siliconflow.com/cn/api-…
3
405
🚀BizyAir's ComfyUI plugin now supports custom LoRA uploads and workflow sharing!
🎉 In just two minutes, you can upload and activate your own LoRA with ease! With cloud storage, it’s faster with less memory.
👉🏻 bilibili.com/video/BV1UHy3YY…
#ComfyUI #FLUX #StableDiffusion
2
354
"A Game-Changing Lib for AI Enthusiasts." Big thanks to @alexgabbia for introducing #OneDiff. #StableDiffusion
🚀 Introducing #OneDiff:
A Game-Changing Lib for AI Enthusiasts.🔥 Unleash unparalleled performance in diffusion models with OneDiff Community Edition. Check these speeds:
-RTX 3090: 42.38it/s 🖥️
-RTX 4090: 74.71it/s 💥
-A100 Variants: Over 50it/s 🌪️
💡 Features:
-Acceleration for popular libraries like ComfyUI, HF diffusers 🤗
-Supports state-of-the-art models including SDXL, LoRA & more.
-Easy drop-in acceleration & multi-res input.
-Plus, enterprise edition offers even more power & flexibility.
🛠️ Easy Install: pip or Docker 👩💻 Open Source & Community Driven 🔗 Dive in: github.com/Oneflow-Inc/onedi…
🌐 Transform your AI journey with OneDiff's unmatched speed & versatility. #AI #SDXL #ComfyUI
2
705
Congratulations to the #InstantID team @Haofan_Wang for their impressive work, which has got 7k stars on GitHub! To accelerate its inference efficiency, #Onediff now achieves 1.8x speedup for InstantID on RTX 4090/3090. Check it out:
teddit.net/r/StableDiffusion…
1
1
2
610
🤖🚀✨ Fresh Models Arrived: OpenAI gpt-oss is now on SiliconFlow! 🔥⚡
This state-of-the-art open-weight model is built for agentic workflows, advanced reasoning, and tool use — with configurable reasoning effort and native support for function calling, web browsing, and Python execution.
🌟 Cost-Effective Pricing
gpt-oss-120B → $0.09 /M input | $0.45 /M output gpt-oss-20B → $0.04 /M input | $0.18 /M output
🌟 Extended Context → 131K tokens
🌟 Configurable reasoning effort – low / medium / high
🌟 Full chain-of-thought access – transparent reasoning
🌟 Fine-tunable – adapt to your use case
🌟 Agentic tool use – function calls, browsing,
Python Benchmarks show gpt-oss-120B outperforms OpenAI’s o3-mini and rivals o4-mini across coding, reasoning, math, health, and tool use. Even the lighter gpt-oss-20B matches or beats o3-mini in several key tasks.
Key innovations include:
🔑 MoE architecture – 117B total params (5.1B active), 21B total (3.6B active)
⚡ Advanced RL alignment – enhances chain-of-thought + tool use
🧩 STEM & coding focus – optimized for real-world developer needs
👉 Try it now：siliconflow.com/models
2
451
🐣The third episode of our BizyAir tutorial series is now available! 🐣
This video dives into understanding LoRA (Low-Rank Adaptation) and guides you through how to upload and use it effectively to enhance your creative projects.
Watch the video here: bilibili.com/video/BV1Dv23Yr…
2
353
Excited to see SiliconFlow API powering multi-provider evaluation and model integration. 🫶🏻
LangGraph + SiliconFlow = smoother multi-model integration & more reliable agent evaluation.
Thanks @zhanghaili0610 for building amazing tools on top of LangGraph!
Upcoming v0.2.0: Multi-Provider Evaluation 🔮
- @SiliconFlowAI: Unified global API for seamless integration of diverse SOTA open-source models.
- Agent Evaluation: Combines OpenEvals and AgentEvals for robust performance testing, leveraging LangSmith as the evaluation tool.
3
413
We support logging in with verification codes sent to overseas phone numbers. You can give it a try!
1
2
38
We are excited to share that #OneDiff has significantly enhanced the performance of SVD (#StableVideoDiffusion by @StabilityAI) on RTX 3090/4090/A10/A100.Especially,OneDiff with DeepCache(@horseeeMa) enables SVD generation speed of up to 3.9x faster #AIGC
teddit.net/r/StableDiffusion…
1
419
FLUX.1 Kontext Dev is now live on SiliconFlow! Big thanks to the BFL team for their amazing contributions to the open-source community! 📷🫡
2
534
👍👍👍
兄弟们，之前薅的 SiliconCloud API 又可以用起来了。
Silo - 纯前端多模型对话、文生图、一对多 AI 模式，响应极快。还支持浏览器插件。通过自定义模型的功能来接入Gemini、Claude、DeepSeek、智谱等。赶紧部署体验一下
github.com/KwokKwok/Silo
1
327
太方便啦！春节马上用上！
拍照查嘌呤的APP「嘌查查」苹果商店限免了。
基于AI识图和语义搜索，效果见视频。
春节饭局多，赶紧转给你身边痛风的同学吧。
支持OpenAI、API2D和国内AI厂商 硅基流动的Key。
推荐使用硅基流动：
语义搜索免费、拍照一分钱 ，注册送2000万Token，能拍一年了… cloud.siliconflow.cn/i/GKAof…
1
963
#InstantID is awesome! Thanks for recommending #OneDiff. It gives you a faster inference experience. Try it out. github.com/siliconflow/onedi…
Thanks @SiliconFlowAI for their OneDiff integration of our InstantID! You can enjoy accelerated inference for InstantID (1.8x acceleration on RTX 4090). You can find more details at github.com/siliconflow/onedi…
1
1
690
必须支持！
大家好，我们开发的麦悠电台上线ProductHunt啦。
它是一个将 RSS 转为 对谈类 Podcast 的iOS应用。和其他类似应用相比，主要是省钱，没有订阅，支持OpenAI/API2d/硅基流动的 API Key，完全客户端加工，支持本地TTS。
求Up求评论呀 → producthunt.com/posts/maidio
1
1,001