Fastest
qwen2.5-0.5b-nnl-distilled-clean-q4_k_m
0.5B
361.0 tok/s
Highest token generation speed at 361.0 tok/s.
Hardware Benchmarks
Best Local LLM Performance for RX 7900 XTX
Compare RX 7900 XTX benchmark results and see which models actually earn the best completed quality scores.
Hardware Overview
GPU-specific specifications for local LLM planning.
- VRAM
- 24GB
- Architecture
- RDNA 3
- Memory Bandwidth
- 960.0 GB/s
Recommended Models
Models we recommend for RX 7900 XTX.
Best quality
unsloth/Qwen3.8-27B-GGUF
27B
37.0 tok/sQuality 91.4
Highest completed quality score at 91.4.
Largest context
gemma4:12b
11.9B
63.1 tok/sContext 196,608 tokens
Largest context that fits in 24GB VRAM, at 196,608 tokens.
Frequently asked questions
Which models actually run best on the RX 7900 XTX, by task, from community benchmark data.
- What is the best local model for coding on the RX 7900 XTX?
- On the RX 7900 XTX, Cyber-Tiel-Coder-35B-A3B-MTP-UD-IQ3_XXS ranks first for coding in our benchmarks (86.3/100 at ~113.8 tok/s). Qwen3.6-35B-A3B-Q4_K_XL is faster (~156.9 tok/s) and still scores well (82.2/100) — a good pick if you'd rather trade a little quality for speed.
- What is the best local model for agentic workflows on the RX 7900 XTX?
- On the RX 7900 XTX, muse-glimmer:latest ranks first for agentic workflows in our benchmarks (93.0/100 at ~34.5 tok/s). unsloth/Qwen3.8-27B-GGUF:Qwen3.8-27B-UD-Q4_K_S is faster (~67.7 tok/s) and still scores well (92.1/100) — a good pick if you'd rather trade a little quality for speed.
Benchmark Results
Completed quality-ranked benchmark results for AMD RX 7900 XTX.