SiliconFlow

💜We are excited to announce that SiliconFlow is now live on OpenRouter!💜 🎯Experience 15+ models, with more being added — including top open ones like Kimi K2, GLM-4.5, DeepSeek V3.1, and others. 🎯Long context, low latency inference, multimodal support — perfect for developers, AI researchers, and those building agents. 🗝️Click to explore & try it out: openrouter.ai/provider/silic… Happy to team up with @openrouter — more to come! 🚀 #AI #OpenSource #AIModels #LLM #developers #SiliconFlow #OpenRouter

Compelling project, agentic storytelling at full scale by agent. Seems like the core runs on Claude Sonnet + Opus for writing, evaluation, and review. Curious if anyone has tried this with open-source models🧐Would love to see how the outputs compare👀

Get your SiliconFlow APIs 👇 cloud.siliconflow.com/account/ak

@kilocode just keeps growing on @OpenRouter — #1 coding agent, 1.5M+ users🧑‍💻 If you're building with it, SiliconFlow integrates nicely since it's OpenAI-compatible with high tool-call accuracy, verified benchmarks, and cost-effectiveness: ➡️Set your base URL to api.siliconflow.cn/v1 ➡️Add one SiliconFlow API key Then, you can freely switch between DeepSeek V3.2 · GLM-5 · GPT-OSS-120B · Kimi K2.5 · MiniMax M2.5 · and more SOTA AI models. Let's get started👇

Folks, if you get crypto emails from websites claiming to be associated with openclaw, it's ALWAYS a scam. We would never do that. The project is open source and non-commercial. Use the official website. Be sceptical of folks trying to build commercial wrappers on top of it.

token ready, waiting for the drop 👀

MiniMax Provider Verifier:github.com/MiniMax-AI/Min…

Get started:siliconflow.com/models/minimax…

Not every MiniMax M2.5 is the same. Out of 6 key metrics, SiliconFlow delivered best-in-class performance in 5.🚀 The latest @MiniMax_AI Provider Verifier proves that SiliconFlow (fp8) isn't just fast and cost-effective, it's also "correct & precise": 🛡️ 100% Query Success Rate: Maximum reliability with a valid response. 🛠️ 84.24% Finish-ToolCalls-Rate: High-performance completion of triggered actions. 🎯 99.28% Trigger Similarity (F1 Score): Decision-making on when to trigger tools, matching official references. 🧠 98.68% ToolCalls Accuracy (Ranked #1): Industry-leading precision for tool-call payloads (function names & arguments). ✅ 100% Response Success Rate - Not Only Reasoning. 🌍 80% Language-Following-success-rate: language requirements in minor language scenarios. In this agent-driven era, Tool Calling precision and its success rate are the difference between a broken workflow and a successful product. ⬇️ Get Started with Verified Excellence ⬇️ Full Benchmark

📎We've uploaded it to arXiv, enjoy! arxiv.org/abs/2603.15031

Thinking how we can evolve openclaw plugins to be more powerful while also making core leaner. Also wanna add support for claude code/codex plugin bundles. Good stuff coming soon!

OpenClaw 2026.3.12 🦞 🎛️ dashboard v2 — slick new control UI ⚡ /fast mode for models 🔌 ollama/sglang/vllm → plugins (core goes on a diet) 🛡️ device tokens now ephemeral ⏰ cron + windows reliability fixes touch grass? nah, touch main 🌿github.com/openclaw/openc…

@openclaw @MiniMax_AI View all models on SiliconFLow and get your API Key today👇siliconflow.com/models

PinchBench benchmarks LLMs as real @openclaw coding agents — scoring each model on success rate, speed, and cost across real-world tasks. In the latest 33-model leaderboard, the most attractive quadrant shows it clearly: open-source models come out on top. @MiniMax_AI -M2.5 · @deepseek_ai -V3.2 · @Kimi_Moonshot -K2.5 · @StepFun_ai -3.5-Flash ... ... with high success rate, low cost. Glad to support these models on SiliconFlow, giving more devs access to SOTA open-source intelligence.

Introducing the new App & Agent Rankings ✨ A better way to explore the AI ecosystem: - Apps can now rank within categories, or make trending - Request-count rankings, in addition to tokens - Apps can easily opt in and categorize their own generations. openrouter.ai/apps

OpenClaw 2026.3.7 🦞 ⚡ GPT-5.4 + Gemini 3.1 Flash-Lite 🤖 ACP bindings survive restarts 🐳 Slim Docker multi-stage builds 🔐 SecretRef for gateway auth 🔌 Pluggable context engines 📸 HEIF image support 💬 Zalo channel fixes We don't do small releases. github.com/openclaw/openc…

Power complex agents without having to worry about cost. @MiniMax_AI M2.5 was built around this idea — frontier-level intelligence at a price most people can actually afford. On SiliconFlow, we've been quietly making sure that promise holds up in practice: 🟢 99%+ uptime this week 💰 $0.078/M input | $1.20/M output — among the lowest across all providers 🎯 82.2% cache hit rate — your repeated calls cost even less MiniMax brings intelligence to everyone. And SiliconFlow scales that reach. Try MiniMax M2.5 on SiliconFlow ➡️ siliconflow.com/models/minimax… You can also use it on @OpenRouter ➡️openrouter.ai/minimax/minima…

"The new version 2.1.68 sets Claude Opus 4.6 to medium effort by default for speed on daily tasks, while typing 'ultrathink' in any prompt triggers high-effort deep reasoning for tough spots."

The Codex app is now on Windows. Get the full Codex app experience on Windows with a native agent sandbox and support for Windows developer environments in PowerShell. developers.openai.com/wendows

Recently, we've received lots of feedback from users asking how to configure SiliconFlow APIs in @openclaw 👀 So here we get a full guide covering: ✅ How to deploy OpenClaw ✅ How to connect SiliconFlow APIs ✅ Step-by-step: adding an OpenClaw bot to your Discord server Check it out 👇 docs.siliconflow.com/en/usercases/h…

We've seen firsthand how much Qwen has meant to developers and builders everywhere — thank you for the incredible work. Wish you the best on the next chapter!