Model benchmarks

qwen/qwen3-4b-2507 local LLM performance

As of November 2026, qwen/qwen3-4b-2507 runs at up to 35.3 tok/s for local inference (best of 4 community benchmark runs across 1 GPU).

LM StudioQ4_K_M
ShareRedditX

Model size

4B

Peak speed

35.3 tok/s

Average speed

34.9 tok/s

Avg PP

1177.9 tok/s

Min memory

4.2 GB

Max context

65,536 tokens

Avg output / run

9,580 tokens

Avg runtime / run

4m 45s

Avg quality

70.1

Benchmark runs

4

GPUs tested

1

Quality by task

Average LLM-judged quality (0–100) with the run-to-run spread shown as a P5–P95 band, overall and for each benchmark task, across all 4 runs. The low and high columns show how much the judge’s score varies between runs, and need at least two runs to display.

TaskP5 (low)AvgP95 (high)
Overall68.270.173.5
Agent Workflow75.386.291.2
Code Generation46.548.850.8
Role Play & Narrative81.483.485.6
Research & Analysis51.062.072.7

Performance by hardware and tool

Every hardware/tool/quantization combination qwen/qwen3-4b-2507 has been benchmarked on, ranked by peak token generation speed. Last updated November 2026.

HardwareToolQuantBest tok/sAvg tok/sMemoryContextQualityRuns
NVIDIA GeForce RTX 2050LM StudioQ4_K_M35.3 tok/s34.9 tok/s4.2 GB65,536 tokens70.14

Benchmark runs

All 4 qwen/qwen3-4b-2507 runs submitted so far — expand one for its hardware, quality breakdown and per-scenario detail.

Frequently asked questions

Is qwen/qwen3-4b-2507 good for coding?
In our benchmarks, qwen/qwen3-4b-2507 scores 48.8/100 for coding. It runs at about 34.9 tok/s, so if you want more speed, byteshape/Qwen3.8-27B-GGUF:IQ4_XS-ASCII-Condensed is faster (~42.0 tok/s) and still scores well for coding (86.3/100).
Is qwen/qwen3-4b-2507 good for agentic (tool-using) tasks?
In our benchmarks, qwen/qwen3-4b-2507 scores 86.2/100 for agentic workflows. It runs at about 34.9 tok/s, so if you want more speed, Nex-N2.5-mini-APEX-Mini is faster (~88.2 tok/s) and still scores well for agentic workflows (91.3/100).
How fast is qwen/qwen3-4b-2507 for local inference?
Across 4 community benchmark runs, qwen/qwen3-4b-2507 reaches up to 35.3 tok/s and averages 34.9 tok/s, with the fastest results on NVIDIA GeForce RTX 2050.
How much memory does qwen/qwen3-4b-2507 need?
The leanest observed configuration used about 4.2 GB of memory (quantizations tested: Q4_K_M).
Which tools have been used to run qwen/qwen3-4b-2507?
Benchmarks were submitted using LM Studio. Results are community-contributed and updated as new runs arrive.