Benchmark result

qwen3.8-27b on NVIDIA GeForce RTX 5070 Ti — 38.6 tok/s

Measured with LM Studio on August 31, 2026.

ShareReddit X

What the model built

Open full screen →

The coding scenario asks for a playable game in a single HTML file. This is exactly what the model returned, unedited.

How this run compares

2nd fastest of 2 runs of this model on NVIDIA GeForce RTX 5070 Ti · median 42.8 tok/s.

sort

146.9tok/s87.8LM Studio
238.6tok/s88.0LM Studiothis run

Compare the nearest 5 →

qwen3.8-27b on other hardware

155.2tok/s89.1NVIDIA GeForce RTX 5090
109.3tok/s86.7NVIDIA L4 +1 more · Q4_K_M
52.4tok/s84.42× AMD Radeon RX 7900 XT/7900 XTX/7900M
52.0tok/s61.72× AMD Radeon RX 7900 XT/7900 XTX/7900M
51.9tok/s86.32× AMD Radeon RX 7900 XT/7900 XTX/7900M

Reproduce this run

toolLM Studio

modelqwen3.8-27b

context121,072

thinkingon

clientv0.4.46+97

Set those in LM Studio, then:

llm-benchmark benchmark --model "qwen3.8-27b" --tool "LM Studio"

The client prompts for context length and thinking mode, and for the KV cache dtype on Unsloth Studio. Sampling is left at the model default — the values above are what the tool reported using, not overrides the client sent.