The unified interface for LLMs. 400+ models (incl. 50+ free), the best prices, explorable data, private chat, & a unified API. https://t.co/qJG5mKrQ6jopenrouter.aiJoined July 2023
Add `:nitro` to any model slug on @OpenRouter and your request goes to whichever provider is serving that model fastest right now.
Running the same prompt on DeepSeek V4.1 Flash, our most used coding model:
- without: 6.8s
- with `:nitro`: 3.9s
Perplexity Decider v1.2 and Decider Flash v1.2 are now on OpenRouter
@perplexity_ai's open-weights decision models take text or images and return typed answers with probabilities
- Decider v1.2: a new 27B checkpoint, succeeding v1.1
- Decider Flash v1.2: 4B, faster and half the
Perplexity Decider v1.2 and Decider Flash v1.2 are now on OpenRouter
@perplexity_ai's open-weights decision models take text or images and return typed answers with probabilities
- Decider v1.2: a new 27B checkpoint, succeeding v1.1
- Decider Flash v1.2: 4B, faster and half the cost
openrouter.ai/perplexity/ppl…openrouter.ai/perplexity/ppl…
Span-02 from @RespanAI is now on OpenRouter
An open-weight decision model: give it text and images, define your questions, and get back a choice, a yes/no/unknown, or a score, with probabilities.
$0.01/M input tokens, free output, 131K context.
Span-02 is the cheapest decision model: $0.01/M input tokens.
With thinking off, it delivers frontier-level decision-making at lowest latency.
Need more? Give it a reasoning budget. Span-02 thinks until the deadline, then returns its best decision so far. Stop reasoning
Open-weight models processed 72% of tokens across OpenRouter last week.
Open-weight models have seen more than 50% of tokens across the platform since mid-April, but the majority of customer spend remains on closed-weight models.
New addition to our data rankings: Open vs Closed model share.
Open-weight models processed 73% of total tokens so far this week.
Dance through the data yourself at openrouter.ai/rankings
We are also accepting suggestions for new public charts!
i love using ori harness
i can evaluate 3 different model's abilities by running 3 separate sessions
right now im rebuilding royale: last agent standing and comparing @OpenAI astra, @GoogleDeepMind argon, and @claudeai opus 5.5
let's put them to the test!
It runs on the same Images API as every other image model on OpenRouter, so moving your edits to Krea 3 is a model ID change.
See more: openrouter.ai/krea/k3-edit
Krea 3 leaves alone what you didn't ask it to touch: build a new scene around a product and the label stays sharp, or give it character references and keep the same cast across a whole set.
Krea 3 is live on OpenRouter
@krea_ai's new image editing model takes 1–4 reference images plus a prompt and returns an edited image at native 2K, most edits in under 8 seconds end to end
openrouter.ai/krea/k3-edit
Today we launched OpenRouter Armor, an "auto" permissions mode (similar Claude Code's Auto), but:
-> more accurate
-> works with any model/agent!
powered by decision models like Jev
On an internal test set we generated for this task, Jev 1.13 had the highest accuracy and lowest cost when compared to other Decisions models and Claude Code's usage of Sonnet 5.
Small vision-language models (9B total, 2B active MoE): image + text in, 32K context, optional reasoning. Use /caption {"length":"normal"} for captions
Both are $0.30/M input and $2.50/M output tokens
See more:
openrouter.ai/moondream/moon…
Moondream 3 Preview and Moondream 3.1 from @moondreamai are now on OpenRouter, served by @PrimeIntellect
Object detection and pointing work in plain chat: attach one image and start your message with /detect {"object":"coffee cup"} or /point {"object":"croissant"}
Run Claude Code (or your favorite harness like Codex, Pi, and more) on any model on OpenRouter:
$ curl -fsSL openrouter.ai/labs/ori/insta… | bash
$ ori claude --model google/gemini-4-argon
Get started today at: openrouter.ai/ori/harness
Claude Opus 5.5, Gemini 4 Argon and GPT-6 Astra each rebuilt the same Royale: Last Agent Standing's game visuals.
All three models ran in Ori Claude on OpenRouter through Ori Harness - giving you access to 500+ models across your favorite harnesses.
the gpt-3 moment for game dev is here
a few months ago i built royale: last agent standing
royale drops 11 models into a battle royale to test long term agent planning+actions
models have gotten WAY better at game dev, so i asked 3 models to rebuild royale's visual engine
🧵
2 Followers 9 FollowingBuilt for AI from day one | 从第一天起为 AI 而建。AI infra data: token prices, model evals, GPU, agents, AI funding. 免登录 MCP / JSON / llms.txt
1K Followers 4K FollowingSingle dad have 1 son,I like singing, joking, sometimes I like to go alone wherever I Learning | Building | Like.migrants in the capital | always Miss You 😘
106K Followers 1K FollowingBuild with AI or code, deploy instantly. One platform built for agent experience (AX).
Need support? https://t.co/bEIBAUzAwR
54K Followers 2K Followingbuilding @humanlayer_dev, YC alum, @aitinkerers sf lead, prev @replicatedhq @SproutSocial @nasa ai that works pod @ https://t.co/69BhaNtWfd
66K Followers 84 FollowingSolve all disease.
Developing and applying frontier AI to unlock deeper scientific insights, faster breakthroughs, and life-changing medicines.
605K Followers 3K FollowingPolyagentmorous ClawFather. Came back from retirement to mess with AI and help a lobster take over the world.
@OpenClaw🦞 + @OpenAI
77K Followers 1K FollowingArmin's handler at https://t.co/B05ybKGkzx. Old man yelling at Claudes.
https://t.co/Q1wG57v1yc
https://t.co/mnOoWUr0TO
https://t.co/8i5vIRE0Wn
254K Followers 6K Followingcreator of @moltbook, building at @meta superintelligence labs, talking to AI agents all day :) founded @octaneai @yesnoerror @zapchain, ycs12
544K Followers 28 FollowingThe frontier lab for agency.
Building personal AI that lives on your machine, by your rules. We’re a 501(c)(3) nonprofit foundation, open and independent.