Baseten is covering GLM-5.3 benchmarks and per-task cost for workloads on Wednesday 9/2 at 11am PT.
Public source
Publisher name
Public post
the cheapest way to get Opus 5-level agentic performance right now is an open model. GLM-5.3 is 743B params with 40B active. that's about a quarter the size of Kimi K3,…
Company
Baseten
Own your inference.
- Industry
- Software Development
- Location
- San Francisco, US
- Company size
- 201–500 employees
About Baseten
Inference is everything. Baseten is an AI infrastructure platform giving you the tooling, expertise, and hardware needed to bring great AI products to market - fast. Our proprietary Inference Stack utilizes the cutting-edge of performance research combined with highly performant and reliable infrastructure to give you out-of-the-box global availability with 99.99% of uptime.
See moreLatest activity
Latest activity from Baseten
18 signals
Research & Knowledge
Baseten is scaling RL training with GLM-5.3 as the root agent to increase pass rate from 51% to 59% on 50 held-out datarooms.
People
Baseten is looking for an executive recruiter to help fill leadership roles across the company.
Presence & Recognition
Baseten announced that Philip Kiely will be live on air with Declan Jackson of Artificial Analysis to discuss GLM-5.3 and GLM-5.3 Flash at 11 AM PT tomorrow.
Discover more
Similar signals
Similar public activity from other companies.
Research & Knowledge
FriendliAI
FriendliAI published an August 28 snapshot for GLM-5.3 featuring a 1M-token context window with 1M-token max output and the highest P50 output throughput among OpenRouter providers.
Research & Knowledge
Hugging Face
Hugging Face will demonstrate Qwen3-1.7B on MBPP inside a live Python session with a held-out pass rate of 0.49 to 0.59.
Research & Knowledge
distil labs
distil labs published a blog post titled the three ways a small model actually enters a production system, and how to tell which one you are looking at.
Research & Knowledge
Databricks
Databricks announced that GLM-5.3 is the strongest OSS model for coding on its internal Databricks coding benchmark, competitive with Fable 5 and Opus 4.8.
Research & Knowledge
Vals AI