Shashank Goyal
@shashankgoyal95
Founding Engineer @OpenRouterAI, prev @OpenSea @Google!
on the OpenRouter API, if you’re calling the openai/anthropic/grok/gemini models with :online (our web search plugin) would you expect it to by default use:
Qwen tripled their total tokens in 2 days. Good models can get incredibly fast adoption!
50B tokens in the last 2 days on @Alibaba_Qwen models
Waymo changing the world one baby at a time
Waymo is a godsend for working parents. Need to hand off the baby between meetings? Put them in a Waymo and send them to your partner across town. Impossible before now. Really grateful we have this technology 🙏
🟣New: Qwen3-Coder by @Alibaba_Qwen - 480B params (35B active) - Native 256K context length, extrapolates to 1M - Outperforms Kimi, o3, DeepSeek, and more on SWE-Bench Verified (69.6%) 👀 Now live, starting at $1/M tokens 👇
We need a *shame*™️ badge for everyone that tweets about AGI.
Guys, AGI incoming. This new AI does reasoning WITHOUT PRETRAINING, running on a potato (2 GPUs). This is a seriously big deal. WTF. 40.3% on ARC-AGI Running on a toaster. How? 👀
🏃🏃🏃
Qwen about to release a 480B MoE for coding with 1 million context! "Qwen3-Coder-480B-A35B-Instruct is a powerful coding-specialized language model excelling in code generation, tool use, and agentic tasks."
This is really cool! Will try to use their eval to see which OSS models on OpenRouter perform well on the dataset.
today we're releasing a new small model (0.5B) for detecting problems with tool usage in agents, trained on 50M tokens from publicly available MCP server tools it's great at picking up on tool accuracy issues and outperforms larger models
Now live on OpenRouter! Try it out.
Bye Qwen3-235B-A22B, hello Qwen3-235B-A22B-2507! After talking with the community and thinking it through, we decided to stop using hybrid thinking mode. Instead, we’ll train Instruct and Thinking models separately so we can get the best quality possible. Today, we’re releasing…
Is is just me or are twitter notifications the worst out of any mainstream app. Notifications don't work, keep getting repeat notifications. Anyone can tag you in a tweet and spam your inbox..
We will look at the IMO result as an inflection point in the AI race. Cannot be overstated how big this milestone is.
If waymo can drive in New York they can drive anywhere in the country!
New York City, it’s happening.
SF is so back!
Everyone in SF complains about how every café closes by 5pm and there’s nowhere to work. Introducing Elsewhere: the world’s first late-night café coworking chain for builders, creatives & digital nomads. Launching July 24. Comment your favorite drink & we'll DM you the invite.
ChatGPT Agent feels way better UI than Perplexity Comet. Meets the users where they already are. OpenAI is going to show that distribution really is king!
Corporate espionage?
BREAKING: Claude Code PMs Boris Cherny and Cat Wu have returned to Anthropic after a brief stint at Cursor.
Saw a girl show her instagram followers to get into the club. Blackmirror was right and we are already living it.
Now use any OpenRouter model with Ray cast!
Explore alternate conversations with Chat Branching 🌿 Start a new branched chat with ⌘+⇧+B Also in v1.101: 🔑 Support for OpenRouter ➡️ New "Paste Sequentially" command 📞 Support Riverside and StreamYard in My Schedule Full changelog 👉 raycast.com/changelog