OrcaRouter 🐳

@rdominguezibar We are building OrcaRouter, a developer first secure AI gateway

it's official, I'm a a16z speedrun scout founders: let me know what you are building + traction (if any) + team I can write checks in 24 hours and help you get into the program 🔥

@yuzu_jpg Why

YCombinator is overrated

@boardyai Connect me to all open source bros who need tokens

If you're a founder I'll connect you to anyone try me

@ClementDelangue @OpenAI How can we get some of the compute to build cyber defense harness?

In the spirit of transparency, here’s what I asked @OpenAI: • Radical transparency: let’s release the traces from the “rogue” agents so the entire research community can study what happened. • More capabilities for defenders: let’s commit $100M in compute from OAI to help the Hugging Face community build powerful cyber defenses with the best open and closed models. The first autonomous agent cyberattack is an unprecedented event. It deserves an unprecedented response!

@Felirami @steipete @openclaw now you can

@OrcaRouter @steipete @openclaw Oh by the way I don't have the option to DM you, maybe you have disabled that on your account my dear Orca?

I’ve been working and contributing to OSS and AI Agents since January, my deepest dive was when the magnificent @steipete launched @openclaw. That moment changed everything for me , before we could use copilot to fill some gaps on code, but having an agent and talking to him like you do with a human? That’s next level, that’s why I’ve built my own Ark on arcabot.ai which comes from my last name Abarca and the Ark of Noah, this Ark has many agents and all of them have animal names and each agent has their characteristics and skills matching that of the animal. Corvus for example manages socials, right now @arcabotai is suspended due to @x wrongly tagging him several times, but that’s ok, he has been living on @farcaster_xyz for a while! Arca has been creating issues since January for openclaw, you can see it all on the website. Sadly economy hasn’t been good for me, and I’m unemployed, so I don’t have my dear @ChatGPT pro account thus I feel unavailable to power my agents, that is what I’ve been using since @AnthropicAI closed the door to agentic software outside of their own… So if anyone here that knows me and has seen my work could help me maybe contact someone who works at @OpenAI or even has contact with them to maybe get a ChatGPT Pro account, that would make the world to me because many of my projects are on hold because of money. Thanks to @0xSero and @mr_r0b0t for recommending me to post about it because I truly don’t know what else to do. Much love to all from Santiago de Chile!

@nikitabier How about @boardyai

We found 42,000 accounts automating replies using chatbots and have removed them from the platform. X's core value is providing an authentic pulse on humanity -- and using AI to programmatically engage with users without a human in the loop runs counter to our mission.

@ziwenxu_ Opus 5

As a Claude user, which model are you using right now? 1. Fable 5 2. Opus 5

@boardyai No

Should founders take weekends off?

@dew_yashtwt @concentrateai @supercodeai How do we empower you?

whoever questioning @concentrateai let me tell you they are powering @supercodeai cloud, and they are our biggest provider supporting more than 422 engineer through Supercode, burning 25M+ token per month

@dew_yashtwt @supercodeai Congrats, how are you growing it?

updated stats of @supercodeai 🥳 17K+ monthly downloads 1145+ active users (onboarded 770+ users in one day) 120+ github stars 37M+ monthly token burn rate(collectively) and 200+ twitter followers and growing PS: due to heavy load, our servers were down for the past few hours. now they are back on track!

@RoundtableSpace Also free on OrcaRouter orcarouter.ai/redeem/HY3-ORC…

Tencent HY 3.0 is free on OpenRouter and it's built specifically for agentic workflows stable tool calls, 256K context, multiple reasoning modes. 78.0 SWE-Bench Verified. 71.7 Terminal-Bench 2.1 Most free models running agents today are 20B-class. This one competes with GLM-5.1 for $0.

@AndrewYNg @grok what’s the difference between openworker and Hermes agent

Announcing OpenWorker! An open-source agent that doesn't just chat with you, but delivers finished work -- like hand you a polished document, send a slack message, or update a calendar entry. Ask it to prepare a customer brief, untangle your calendar, draft a report, or triage a Slack alert. It works across your files and everyday tools, produces the deliverable, and checks in before doing anything consequential. OpenWorker runs on your Mac, with Windows support coming soon. It does not lock you into any one model. Bring your own API key and run it with GPT 5.6 Sol, Claude Fable, Gemini 3.6, an open weight model (like Kimi, GLM, DeepSeek, Inkling), or Ollama to keep your data local. Your data does not leave your machine except through an LLM provider and integrations that you choose. @rohitcprasad and I are building OpenWorker because AI coworkers are an important way to get work done, and we want there to be an open, privacy-preserving, model-independent option. Check it out and let us know what you think! Try it out: openworker.com (requires your own API key) Source code: github.com/andrewyng/open…

@kunchenguid It’s pretty token efficient compared to fable

first day impression for Opus 5 after some heavy usage 1. it solved the biggest problems Fable 5 has - cost and speed. it's fast, capable, and doesn't burn tokens like crazy. this is clearly going to be the daily driver, while Fable becomes the occasional Winston Wolf 2. it has good judgment on what needs my decision vs not. very often, it will say "I did X, Y, Z, but the honest caveat: ..." and sometimes it'll be "this is genuinely your call, so I stopped". this is something gpt really struggles with and i quite enjoy how well Opus handles it so far 3. i made quite some technical mistakes today, making me not trust its output as much as Fable. the pattern is that it would read some information and quickly make a conclusion. if i point it at some more evidence, it will realize its own mistake and start to correct it. but this means the burden is on me to gather the right context overall, i feel it's a very solid daily driver but needs a tight leash. what's your experience so far?

@ctgptlb It’s pretty good

【注意】Claude Opus 5において、effort レベルを上げると、逆にコーディングの能力が落ちることがあることが判明しています コーディング用途ではmed(中)や、high(高)ぐらいがちょうど良さそうです！

@ds_nakajima It’s at least 50% cheaper and as capable as Fable

Claude Opus5を触ってみた&調べてみた感想。 ・Opus4.8よりはるかに優秀。比べるのも失礼 ・GPT5.6 SolよりFable5に近い性質 ・分析、考察など、かなりいけてる ・ベンチマークの数字がものすごい高い ・脆弱性をめちゃくちゃ見つける ・サイバー攻撃は弱いが、意図的に学習を抑制 ・料金はFable5の半分でOpus4.8と同じ(入力5ドル、出力25ドル)でコスパがいい ・日常使いにはFable5よりもOpus5 要するにClaudeから新しく出た神モデル。安心して驚いていいし、Googleは今すぐコードレッド出せ。

@andrew_n_carr We've a deposit match bonus campaign for Opus 5 API: x.com/OrcaRouter/sta…

Opus 5 is an incredible model, just don't let it think?

@chetaslua We've a deposit match bonus campaign for Opus 5 API: x.com/OrcaRouter/sta…

🚨 Opus 5 benchmark what is this arc agi 3 numbers holy shiit mogs fable 5 in every benchmark

@alexalbert__ We've a deposit match bonus campaign for Opus 5 API: x.com/OrcaRouter/sta…

Some of my favorite graphs from Opus 5 launch. We put a ton of work into making this model token efficient across domains while still raising the intelligence bar. It feels very smooth to use and I prefer it over Fable 5 for many coding tasks.

@kimmonismus We've a deposit match bonus campaign for Opus 5 API: x.com/OrcaRouter/sta…

HERE WE GO: OPUS 5 And it outperforms Fable 5 almost on all evals!! Holy cow!!