All model and GPU pairings
GLM 5.3 744BNVIDIA Blackwell

Running GLM-5.2 on GB300 NVL72

Quick answer

GLM-5.2 runs on GB300 NVL72: 14 benchmarked configs so far. See the interactivity ladder below for measured operating points.

Benchmarked configs

14

Serving engines

Dynamo SGLang, Dynamo TRTLLM

Precisions

fp4

Run dates

2026-08-162026-08-21

Throughput at every interactivity target

Serving is a trade-off: push more concurrent users through a GPU and each user's tokens arrive slower. The ladder below reads the measured frontier at each per-user speed target on the AgentX agentic coding workload, using the best engine and precision at that point.

Per-user targetTokens/s per GPU$ / 1M tokensEnginePrecision
30 tok/s--Dynamo SGLangfp4
50 tok/s--Dynamo SGLangfp4
75 tok/s--Dynamo SGLangfp4
100 tok/s--Dynamo SGLangfp4
150 tok/s9,497$0.068Dynamo SGLangfp4
200 tok/s--Dynamo SGLangfp4

Frequently asked questions

How fast is GLM-5.2 on GB300 NVL72?
The InferenceX fleet has 14 benchmarked configs for this pairing; see the interactivity ladder above for the operating points reached so far.
How much does it cost to serve GLM-5.2 on GB300 NVL72?
Cost per million tokens is derived from measured throughput and $/GPU/hr rates from the SemiAnalysis AI Cloud TCO model; it appears once this pairing reaches the primary interactivity tier.
Which serving engines run GLM-5.2 on GB300 NVL72?
The runs behind this page used Dynamo SGLang, Dynamo TRTLLM in FP4, including disaggregated prefill and multi-node serving. Engines are rebuilt and re-benchmarked continuously, so the best config can change between visits.
How are these GLM-5.2 numbers measured?
Every number is measured on real GB300 NVL72 hardware by the InferenceX fleet, sweeping concurrency on the AgentX agentic coding workload to trace the throughput-versus-interactivity frontier; the newest run landed on 2026-08-21. The same derivation powers the InferenceX overview leaderboard.

Explore the data