Benchmark result
llama-3.2-3b-instruct on NVIDIA GeForce RTX 2050 — 45.6 tok/s
Measured with LM Studio on October 20, 2026.
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; the frame only adds a line that tells this page how tall it is.
How this run compares
4th fastest of 4 runs of this model on NVIDIA GeForce RTX 2050 · median 45.7 tok/s.
sort
- 145.9tok/s32.4LM Studio · Q4_K_M
- 245.7tok/s36.7LM Studio · Q4_K_M
- 345.7tok/s36.0LM Studio · Q4_K_M
- 445.6tok/s32.3LM Studio · Q4_K_Mthis run
Reproduce this run
toolLM Studio
modelllama-3.2-3b-instruct (Q4_K_M)
context8,192
clientv0.4.96+102
Set those in LM Studio, then:
llm-benchmark benchmark --model "llama-3.2-3b-instruct" --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.