Best local LLM for Spanish

Measured Spanish scores of quantized local models, per GPU. Highest average so far: qwen3.8-27b-q3 on AMD Radeon RX 9070 XT.

Model	GPU	Decode tok/s	Knowledge (translated) %	Local exams %	Math %	Translation chrF++	Source
`qwen3.8-27b-q3`	AMD Radeon RX 9070 XT	50	80.2	76.8	-	-	maintainers
`gemma4-26b-a4b-qat`	AMD Radeon RX 9070 XT	90	74.5	75.2	-	-	maintainers
`qwen3.5-9b-q4`	AMD Radeon RX 9070 XT	42	76.0	73.2	90.0	56.5	maintainers
`qwen3.5-4b-q4`	AMD Radeon RX 9070 XT	78	70.5	70.0	88.8	55.1	maintainers
`gemma4-e4b-q4`	AMD Radeon RX 9070 XT	105	66.2	70.8	88.8	55.6	maintainers
`qwen3.8-27b-iq2`	AMD Radeon RX 9070 XT	40	70.8	69.2	-	-	maintainers

Zero-shot, thinking off; margins about ±4-5 points. How each column is measured: leaderboard. Missing your GPU? Run localllm eval --langs en,es and share it with localllm report.