LLM Inference Benchmark Explorer

Kimi K3 on DGX B300 — MXFP4, vLLM, TP8, speculative decoding inference benchmark

Open this configuration in the LLM Inference Benchmark Explorer

ModelKimi K3
Parameters2.78T
Intelligence Index43.6
Agentic Index50
DeviceDGX B300
QuantizationMXFP4
Max C32
Chat Capacity128
Agentic Capacity48
TP8
DP—
PP—
EnginevLLM
Speculative DecodingYes
CTTFT (ms)TPS (tok/s)Status
171187.74PASS
2173129.75PASS
419199.43PASS
826361.04PASS
1634443.60PASS
3256523.14PASS
64103111.21FAIL

In OpenZeka's measurement, Kimi K3 (2.78T parameters), served in MXFP4 format with vLLM and speculative decoding on DGX B300 (TP=8), reached a generation speed of 187.7 tok/s per request and a time to first token (TTFT) of 71 ms with a single request.

Considering the speed targets and the available KV cache capacity, the estimated capacity is 128 users for chat use and 48 for agentic use. The realistic capacity will likely fall between these two values. In scenarios dominated by coding, tool use, long workflows and multi-agent use, capacity approaches the agentic estimate; where shorter interactions, standard conversations and lighter tasks dominate, it approaches the chat estimate.

Notes: DSpark speculative. Draft model: Inferact/Kimi-K3-DSpark. 1.86x speedup at C1.

Intelligence Index and Agentic Index values are published by Artificial Analysis and are reproduced here with attribution.