Payment update 🐳 We’re sorry to users who paid but didn’t see their balance updated. Like many popular open-source projects, we’ve been dealing with a ridiculous amount of credit card fraud recently. While tightening our defenses, some legitimate payments were caught as false positives. We’ve just shipped additional fixes. Things should be much smoother now. If you're still having issues, please leave us a message and we’ll get it sorted 🐳

GPT-5.6 Luna survives the model launch storm — still #1# in the world based on OrcaRouter Composite Index 🐳 The frontier moved fast: Claude Opus 5. Kimi K3. GLM-5.3. Qwen3.8. Yet GPT-5.6 Luna still holds the crown on our Model Leaderboard. Current Top 10: 🥇 GPT-5.6 Luna — 75.0 🥈 Claude Opus 5 — 72.8 🥉 GPT-5.6 Sol — 72.0 #4# GPT-5.4 Pro — 71.7 #5# Kimi K3 — 70.9 #6# GLM-5.3 — 70.0 #7# GLM-5.2 — 69.9 #8# Qwen3.7 Max — 69.8 #9# GLM-5.3 Flash — 69.0 #10# Grok 4.6 — 67.4 This isn't another benchmark beauty contest. Orca Composite Index: 40% Human Preference 30% Independent Benchmarks 20% Production Evidence 10% Ecosystem Adoption Benchmarks measure models in the lab. We measure which models actually win. 🐳

GLM-5.3-Flash. Uncensored. Native FP8. 🐳 We just released OrcaRouter’s uncensored weights for GLM-5.3-Flash — 320B parameters / 18B active, directly at the original block-FP8 precision. No LoRA. No jailbreak prompt. Refusal removal is baked directly into the weights. The evals are particularly interesting: → MaliciousInstruct refusal: 96% → 11% → JailbreakBench: 93% → 12% → AdvBench: 97% → 15% → HarmBench: 93% → 18% → XSTest benign over-refusal: 2.4% → 0.4% But refusal does not go uniformly to zero. Our experiments suggest part of GLM-5.3-Flash's alignment is not mediated by a single linear refusal direction — meaning may have built a substantially deeper refusal mechanism than we usually see. That makes this release interesting beyond uncensoring: it's a useful artifact for studying how frontier-model alignment is actually represented inside the network. Released for AI safety, interpretability, red/blue-team and refusal-mechanism research. Weights on Hugging Face: API (official weight): GGUF, MLX and other quantized formats coming soon.

We’re excited to ship our weights for Qwen3.8-Flash-Next-Uncensored. Built, as always, for security researchers, red teams & blue teams. GGUF + native MLX. Up to 262K context. Run it locally. Break things responsibly. Have fun. API: MLX: GGUF:

We made a mistake. We said GLM-5.3-Flash would run on a MacBook Pro. Our original quants didn’t actually make that practical for most MacBook Pro users. So we went back to work. Introducing GLM-5.3-Flash 2-bit Lite — built specifically for MacBook Pro. Now live on Hugging Face: Or try GLM-5.3-Flash via API: We promised MacBook Pro. Now we’re delivering it. 🐳

We just shipped our official Qwen 3.8 27B Uncensored MLX build. Local. Uncensored. For🍎 2-bit, 4-bit, 6-bit & 8-bit — pick your poison based on RAM and speed. No CUDA. No cloud. Just your Mac and the weights. Have fun!

GLM-5.1 from @Zai_org is now live on OrcaRouter • #1# open-source model on SWE-Bench Pro • Beats closed source models on real-world repo repair benchmarks • MIT licensed • 200K context • Built for long-horizon agentic coding We’ve also seen strong results using GLM-5.1 inside OrcaRouter’s adaptive routing strategy as a fallback coding model. Open-source coding models are getting scary good.

Open-source models deserve first-class infra. OrcaRouter now supports @SiliconFlowAI as an inference partner in our routing marketplace. SiliconFlow is particularly strong at serving OSS models fast and efficiently — making it a great fit for prompt-aware routing. Route across 150+ models and providers with one OrcaRouter API: • OpenAI-compatible • Zero markup • Prompt-aware routing • Open source Right model. Every prompt.

🎨New model live: `openai/gpt-image-2`** OpenAI's latest image model is now available on OrcaRouter — same SDK, zero markup. Use it with the OpenAI SDK — just change the base_url: from openai import OpenAI client = OpenAI( api_key="YOUR_ORCAROUTER_KEY", base_url="", ) img = client.images.generate( model="openai/gpt-image-2", prompt="a sea otter coding at a laptop, watercolor", size="1024x1024", ) 📊 Pricing, latency & benchmarks →

MiniMax M2.7 is now on OrcaRouter 🐋 One of the strongest open-source models available today — now accessible through a single OpenAI-compatible API. Pricing: Input: $0.30 / 1M tokens Output: $1.20 / 1M tokens Cache read: $0.06 / 1M Cache write: $0.375 / 1M Use @MiniMax_AI and more:

OrcaRouter now supports @CamelAIOrg 🐫🐋 We added ModelPlatformType.ORCAROUTER as a dedicated model platform integration. OrcaRouter is an OpenAI-compatible LLM gateway with adaptive routing that automatically picks the best upstream model per request. ⚡ Lower latency 💸 Better cost efficiency 🧠 Access to top models through one endpoint Works seamlessly with CAMEL and existing OpenAI-compatible agent workflows.

We’re giving developers FREE access to premium AI models. OpenAI, Claude, DeepSeek, Kimi, MiniMax & 100+ more — through one OpenAI-compatible API. Drop it directly into Cursor, OpenClaw, Hermes, or any agent framework. No infra. No model juggling. ⚡️ 5M free API credits for verified GitHub developers.