Introducing Replicas V3: the most flexible and cost-efficient cloud agent on the market.
- Run Claude Code, Codex inside VMs
- Bring any subscription or API keys
- Delegate from Slack, Linear, desktop app, or your phone
Now proactive, and more powerful than ever.
replicas.dev
is it a pmf signal if a company is a top 10 user of your product, inflates the hell out of your run rate to the point where the team is actively being freaked out by the "sales-gong" channel, and won't respond to your emails
we've onboarded a bunch of new teams in the last week, almost all using cloud agents for the first time
by far the best predictor of ROI for them is kindness, because it directly correlates with how easily they mold their dev workflows to be more efficient, and ask us great questions (and in turn its easier for us to best serve them!)
i've been taking this approach with our vendors - turns out people who work in AEO 24/7 may, in fact, know more about AEO
📢 Meet Qwen4-Max, your new AI teammate for work, creativity, and discovery!
Qwen4-Max is built on a new MoE architecture with native vision and a 1M-token context window.
Highlights:
- Outperforms Qwen3.8-Max with up to ~50% fewer tokens, while average cost per task drops over 60% in our internal evaluation.
- Advances real-world productivity through greater proficiency with professional tools, stronger long-horizon execution, and higher-quality deliverables.
- Handles real tasks, including 3D modeling, game development, manufacturing, finance, law, healthcare, and frontier scientific discovery.
- $1.5/M tokens for input, $4.5/M tokens for output, and $0.075/M tokens for implicit caching — all lower than Qwen3.8-Max, with implicit caching down by 70%.
The API is live on QwenCloud. You can also try Qwen4-Max on Qwen Studio, QwenWork and Qoder.
Start building today! 🚀
- Blog: qwen.ai/blog?id=qwen4-max
- API: qwencloud.com/models/qwen4-m…
- Token Plan: qwencloud.com/pricing/token-…
More to come! The open weights of smaller Qwen4 models will be released. Stay tuned. 👀
we had a noticeable drop in usage a little after Demo Day in June (you can see it), it honestly really scared me and my cofoudner
but it gets less and less noticeable every day on the y-axis, which is a fantastic feeling
We built an inference engine for DeepSeek V4.1-Flash from scratch in Rust and CUDA.
4xB300: ~1,740 TPS decode
8xB300: ~2,060 TPS decode
1M-token prefill in ~12s
It's open-source.
yeah I mean tiny sample size, maybe I got lucky and met the nicest most motivated purdue guys
we literally just hired Waterloo so don’t overindex on my ramble tweet 💀
on our offsite our engineer told us that support requests per customer have gone way down but we’re STILL aggressively hiring because there’s that much eng work
idk what we’d do w/o Saai cooking in this
The replicas.dev codebase is now self maintaining. All customer requests and bug fixes happen proactively without an engineering prompting an agent.
Here's how we do it