Model benchmarks

turbo local LLM performance

As of November 2026, turbo runs at up to 135.6 tok/s for local inference (best of 2 community benchmark runs across 1 GPU).

radiance
ShareRedditX

Model size

Unknown

Peak speed

135.6 tok/s

Average speed

124.1 tok/s

Avg PP

3419.5 tok/s

Min memory

n/a

Max context

65,536 tokens

Avg output / run

77,357 tokens

Avg runtime / run

10m 7s

Avg quality

89.4

Benchmark runs

2

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 2 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)
Overall89.489.489.4
Agent Workflow84.087.591.0
Code Generation81.685.389.0
Role Play & Narrative92.094.196.2
Research & Analysis89.090.892.6

Performance by hardware and tool

Every hardware/tool/quantization combination turbo has been benchmarked on, ranked by peak token generation speed. Last updated November 2026.

HardwareToolQuantBest tok/sAvg tok/sMemoryContextQualityRuns
AMD Radeon AI PRO R9700radiance—135.6 tok/s124.1 tok/sn/a65,536 tokens89.42

Benchmark runs

All 2 turbo runs submitted so far — expand one for its hardware, quality breakdown and per-scenario detail.

Frequently asked questions

Is turbo good for coding?
In our benchmarks, turbo scores 85.3/100 for coding. It runs at about 124.1 tok/s, so if you want more speed, ornith-1.5-35b-apex is faster (~149.2 tok/s) and still scores well for coding (83.5/100).
Is turbo good for agentic (tool-using) tasks?
In our benchmarks, turbo scores 87.5/100 for agentic workflows. It runs at about 124.1 tok/s, so if you want more speed, swift-1.5-qwen3.8-flash-next-iq3_s is faster (~259.0 tok/s) and still scores well for agentic workflows (88.9/100).
How fast is turbo for local inference?
Across 2 community benchmark runs, turbo reaches up to 135.6 tok/s and averages 124.1 tok/s, with the fastest results on AMD Radeon AI PRO R9700.
Which tools have been used to run turbo?
Benchmarks were submitted using radiance. Results are community-contributed and updated as new runs arrive.