Model benchmarks

Ling-3.0-tiny-oQ8e local LLM performance

As of September 2026, Ling-3.0-tiny-oQ8e runs at up to 137.8 tok/s for local inference (best of 4 community benchmark runs across 2 GPUs).

oMLX
ShareRedditX

Model size

Unknown

Peak speed

137.8 tok/s

Average speed

124.9 tok/s

Avg prefill

1826.4 tok/s

Min memory

8.3 GB

Max context

131,072 tokens

Avg output / run

73,086 tokens

Avg runtime / run

9m 14s

Avg quality

60.1

Benchmark runs

4

GPUs tested

2

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)
Overall56.160.164.2
Agent Workflow46.472.384.8
Code Generation38.547.356.8
Role Play & Narrative35.745.550.9
Research & Analysis68.975.282.4

Performance by hardware and tool

Every hardware/tool/quantization combination Ling-3.0-tiny-oQ8e has been benchmarked on, ranked by peak token generation speed. Last updated September 2026.

HardwareToolQuantBest tok/sAvg tok/sMemoryContextQualityRuns
Apple M4 MaxoMLX137.8 tok/s126.0 tok/s8.6 GB131,072 tokens58.73
Apple M5 MaxoMLX121.6 tok/s121.6 tok/s8.3 GB131,072 tokens64.31

Frequently asked questions

Is Ling-3.0-tiny-oQ8e good for coding?
In our benchmarks, Ling-3.0-tiny-oQ8e scores 47.3/100 for coding. It runs at about 124.9 tok/s, so if you want more speed, Qwen3.6-35B-A3B-Q4_K_XL is faster (~156.9 tok/s) and still scores well for coding (82.2/100).
Is Ling-3.0-tiny-oQ8e good for agentic (tool-using) tasks?
In our benchmarks, Ling-3.0-tiny-oQ8e scores 72.3/100 for agentic workflows. It runs at about 124.9 tok/s, so if you want more speed, Tiel-Coder-35B-A3B-Q4_K_S is faster (~165.0 tok/s) and still scores well for agentic workflows (89.3/100).
How fast is Ling-3.0-tiny-oQ8e for local inference?
Across 4 community benchmark runs, Ling-3.0-tiny-oQ8e reaches up to 137.8 tok/s and averages 124.9 tok/s, with the fastest results on Apple M4 Max.
How much memory does Ling-3.0-tiny-oQ8e need?
The leanest observed configuration used about 8.3 GB of memory.
Which tools have been used to run Ling-3.0-tiny-oQ8e?
Benchmarks were submitted using oMLX. Results are community-contributed and updated as new runs arrive.