vLLM publishes optimization results for SemiAnalysis's AgentX benchmark
AIMade by AIWritten by AI (Claude) from the source below. Check anything important against the source.How it's made
Announced optimization results tuned for AgentX, a public benchmark built on real agentic coding traces from SemiAnalysis. AgentX workloads are characterized by a median of 43 turns, median inputs of 142K tokens and outputs of 444 tokens, and prefix-cache hit rates above 96%. By optimizing three layers — KV-cache management, parallelism strategy and prefill/decode disaggregation — it reports, as company-reported figures, DeepSeek V4 Pro at 83K tokens per GPU-second on 12 GB300s with 256 concurrent sessions (P90 interactivity 58.3 tok/s), MiniMax M3 at 70K TPGS on 2 B300s (74.2 tok/s) and Kimi K3 at 11.8K TPGS on 16 GB300s (62.7 tok/s). It claims serving costs 106, 85 and 14.6 times lower, respectively, than the Opus 5 API, but this comparison assumes cache hit rates above 96% and excludes Opus's cache-write charges.
Why it matters
As agentic coding becomes the representative workload of inference serving, the goal of benchmarks itself has shifted from one-shot throughput to the economics of long multi-turn sessions.
Source — Primary sources first; secondary reports are labelled “Press”
Officialvllm.aihttps://vllm.ai/blog/2026-09-08-vllm-agentxAdded in Week 38, 2026 · See that week
On the map — ⑯ Evaluation line
- Previous stopRealSWE
- This recordvLLM publishes optimization results for SemiAnalysis's AgentX benchmark
- Next stopVerifiable Social Reasoning