Baseten

@baseten

Inference is everything.

San Francisco and New York

baseten.co

Replies Reposts Media Articles

Pinned
Baseten
@baseten
Sep 3
We’re excited to keep pushing the open-source AI ecosystem forward.
Base Labs
@baselabs
Sep 3
Today we're announcing Base Labs, a dedicated research organization focused on advancing open-source AI. We believe in a healthy, open frontier model ecosystem. To enable this, we are working on: - Blue-sky research on continual learning, the science of RL, and how models learn
00:00
3
Baseten
@baseten
11h
We're excited to bring Mercury 2.5 to Baseten: - Sub-200ms p50 latency, great for real-time voice use cases - 1,100+ tok/s on @nvidia_ai GPUs Get access through our model library: baseten.co/library/mercur…
Inception
@_inception_ai
18h
Today, we’re introducing Mercury 2.5, the most capable diffusion LLM on the market. It offers a 40% jump in intelligence over Mercury 2, and runs over 1,100 tokens/sec on widely-available @NVIDIAAI GPUs. It’s available today on our API, @OpenRouter, and @baseten. Contact us
00:00
3
Baseten
@baseten
13h
We're proud to define the quality-latency Pareto frontier for STT in Coval's benchmarks. Voice AI is an inference problem, and teams need benchmarks that measure inference to know what their users will actually feel. Proud to partner with @covaldev for an open, reproducible
Baseten leads Coval’s voice AI benchmark
From baseten.co
2
Baseten
@baseten
16h
AI labs and frontier research teams run RL managed rollouts on Baseten. They publish new weights, and we pick them up across a global fleet so rollouts keep generating against the latest policy. On average, a new policy is live in under 40 seconds with just a 6-second request
Paras Stefanopoulos
@stefanopopoulos
16h
Sub 40s delta weight syncs for GLM 5.3 🔥💚
Article
Optimizing Delta Weight Syncs for Managed Rollouts
Baseten supports delta weight syncs to leading frontier models like GLM-5.3 across global, independent clusters in under 40 seconds, with only 6 seconds of request pause. AI labs and frontier research...
Baseten
@baseten
17h
We partnered with @harvey to train agents for M&A diligence and showed that model-harness co-optimization can improve agent performance. By applying RL to a Qwen 122B-A10B root agent inside a RLM harness for M&A diligence, we demonstrated pass rate improvement from 29% to 63%.
Niko
@nikogrupen
18h
Article
Post-training RLM agents for end-to-end M&A Diligence
Article link: https://www.harvey.ai/blog/post-training-rlm-agents-for-m-and-a-diligence In our recent research preview for Harvey Tenet, we highlighted the importance of model-harness co-optimization for solving complex, end-to-end legal tasks, and...
1