We just shipped a number of updates to the OpenRouter MCP - 2 weeks of OpenRouter MCP updates in one thread + demo.
Your agent can now discover, rank, test, and report on models without leaving your editor 🧵
Two new tools: generate-image (works on Desktop apps that can view images, billed to your key) and send-feedback (report a bad generation straight to the OpenRouter team, no support form).
Plus: docs search now runs on @mintlify, read-only tools no longer trigger permission prompts, and every tool renamed to a clean verb-noun format. Get started in 30 seconds:
New favorite Codex workflow for non-technical work:
"Please use @OpenRouter to ask 3-4 other smart models to challenge your conclusions, and to help decide on the correct answer, update your answer accordingly"
10 weeks since launch.
@poolsideai's Laguna M.1 is #11 among open models and #18 across ALL models on @OpenRouter by monthly usage, holding rank in one of the fastest-moving leaderboards in AI. 🔥
Just getting started.
Service tiers are now first-class on OpenRouter.
See tier-specific latency and throughput in `/v1/models/endpoints` so you know exactly what you’re paying for.
Route with copyable slugs like `openai/flex` and `xai/priority`, or keep using the existing `service_tier` param.
hot take: choosing the best coding model is easy
but for everything else - translations, summarization, data extraction, etc - it's all based on vibes
we just shipped task type insights on @OpenRouter mcp - see what models ppl are using for specific tasks, backed by prod usage
OpenRouter has a new look.
As we’ve grown into the infrastructure powering the intelligence era, our brand needed to evolve too.
Built on clarity, precision, and purpose.
For developers. For enterprises. For everyone turning intelligence into impact.
GPT-5.6 is live on OpenRouter!
Sol (flagship), Terra (balanced), and Luna (fast + low-cost).
In our early testing, the 5.6 models were more token efficient, leading to lower cost and faster time to task completion. AA's results agree.
More on Sol, Terra, and Luna below 🧵
GPT-5.6 Luna: the fast, low-cost tier. Tuned for high-volume, latency-sensitive work like chat, classification, and lightweight agents, with real GPT-5.6 reasoning.
Announcing Grok 4.5, our first model trained specifically for coding and agents. It was trained with Cursor and offers frontier intelligence at leading speeds and cost efficiency.
x.ai/news/grok-4-5