the reasoning age begins soon.

l1 is out—and the world noticed. when we open-sourced it, we didn’t run ads. no launch event. no hype train. just a simple model, built to be fast, light, and open to everyone. we pushed it to hugging face, dropped the repo on github, and let the code speak. within days, people were building with it. we saw clones, ports, live demos. devs started dropping it into their pipelines and tagging us in screenshots. we expected curiosity. we got energy. it’s not just validation—it’s ignition. me, ujjwal, and firoj (our newly minted chief research officer) are in it. we’re moving fast, but the clarity is even faster. we wrapped our site last night—it’s clean, bold, and drops next week. our api infra is almost already running, serving up inference behind the scenes while we harden it for public launch. we’re laying bricks like we’ve done this a hundred times, but everything still feels like a first. v1 is on deck—our flagship model, bigger, sharper, and still cut from the same cloth: speed, efficiency, and no wasted motion. it’s designed to do heavy lifting without dragging the system down. and it’s not a solo act. because sage is coming. mid-june, if we hit our timeline. we’re building sage to reason—not just answer. it’s lean, it’s deliberate, and it feels like it’s thinking with you. not bloated. not noisy. just clean logic, stripped to the core. we’re testing it hard. it’s outperforming where it counts. and it’s got us up at night, excited like we were back in month one. we haven’t shown much yet. for good reason. sage is the heart of this next wave—a compact reasoning model designed to hit with the clarity and depth of models ten times its size. it doesn’t just generate—it breaks things down. it doesn’t just respond—it reasons through. we’re pressure-testing it against tasks that most models fake their way through. and so far, it’s holding up. there’s something different about this run. maybe it’s the feedback loop from the community, or the way l1’s already showing up in indie projects. maybe it’s the clarity we’ve found—build only what matters, cut the rest. sagea’s still small, but it’s thinking big. we’re not here to imitate. we’re here to carve. we’re tuning every model like it’s a machine with soul. voice, vision, language, and now: reasoning. not just ai that responds. ai that engages. our roadmap’s not packed—it’s precise. and that’s why it’s working. if you’ve built with l1, thank you. if you haven’t, you should. the site goes live next week. v1’s coming soon. and sage… sage is nearly awake. — ciao, basab