What is the best local model for coding on the 2× RX 9050 9060 Xt?

On the 2× RX 9050 9060 Xt, qwen3.8:27b-mtp-q4_K_M ranks first for coding in our benchmarks (82.8/100 at ~24.2 tok/s).

What is the best local model for agentic workflows on the 2× RX 9050 9060 Xt?

On the 2× RX 9050 9060 Xt, qwen3.8:27b-mtp-q4_K_M ranks first for agentic workflows in our benchmarks (87.3/100 at ~24.2 tok/s). gemma4:12b-it-qat is faster (~33.5 tok/s) and still scores well (82.5/100) — a good pick if you'd rather trade a little quality for speed.

Hardware Benchmarks

Best Local LLM Performance for 2× RX 9050 9060 Xt

Compare 2× RX 9050 9060 Xt benchmark results and see which models actually earn the best completed quality scores.

ShareReddit X

Hardware Overview

GPU-specific specifications for local LLM planning.

2× AMD Radeon RX 9050 / 9060 XT

VRAM: 30GB
Architecture: Multi-GPU

Recommended Models

Models we recommend for 2× RX 9050 9060 Xt.

Fastest

gemma4:12b-it-qat

11.9B

33.5 tok/s

Highest token generation speed at 33.5 tok/s.

Best quality & Largest context

qwen3.8:27b-mtp-q4_K_M

27.3B

24.2 tok/sQuality 88.1Context 65,536 tokens

Highest completed quality score at 88.1. Largest context that fits in 30GB VRAM, at 65,536 tokens.

Frequently asked questions

Which models actually run best on the 2× RX 9050 9060 Xt, by task, from community benchmark data.

What is the best local model for coding on the 2× RX 9050 9060 Xt?: On the 2× RX 9050 9060 Xt, qwen3.8:27b-mtp-q4_K_M ranks first for coding in our benchmarks (82.8/100 at ~24.2 tok/s).
What is the best local model for agentic workflows on the 2× RX 9050 9060 Xt?: On the 2× RX 9050 9060 Xt, qwen3.8:27b-mtp-q4_K_M ranks first for agentic workflows in our benchmarks (87.3/100 at ~24.2 tok/s). gemma4:12b-it-qat is faster (~33.5 tok/s) and still scores well (82.5/100) — a good pick if you'd rather trade a little quality for speed.

Benchmark Results

Completed quality-ranked benchmark results for 2× AMD Radeon RX 9050 / 9060 XT.

Open run page →