# Radeon RX 7900 XTX 24 GB

> 24 GB for less than a used 3090 at some retailers; ROCm and Vulkan backends both work in llama.cpp.

Canonical: https://thelocalbench.com/hardware/rx-7900-xtx-24g · Data as of 2026-10-12 · The Local Bench (thelocalbench.com), data as of 2026-10-12, CC BY 4.0

Flags: [verified] checked against primary source · [sourced] cited · [estimate] formula · [sample] placeholder

## Specifications

| Field | Value | Flag | Source / note |
|---|---|---|---|
| memory_gb | 24 | sourced | https://www.amd.com/en/products/graphics/desktops/radeon/7000-series/amd-radeon-rx-7900xtx.html |
| bandwidth_gbs | 960 | sourced | https://www.amd.com/en/products/graphics/desktops/radeon/7000-series/amd-radeon-rx-7900xtx.html |
| fp16_tflops | 123 | sample | Sample FP16 dense compute; pending vendor-sheet citation. |
| load_watts | 400 | sample | Sample wall power under load. |
| price_usd | 899 | sample | Sample price. |

## Models that run (default quant, 8k)

| Model | Quant | Memory GB | Verdict | Decode tok/s |
|---|---|---|---|---|
| [Llama 3.1 8B Instruct](/models/llama-3.1-8b.md) | Q4_K_M | 6.6 | fit | 96.1 [estimate] |
| [Qwen3 8B](/models/qwen3-8b.md) | Q4_K_M | 6.8 | fit | 92.3 [estimate] |
| [Qwen3 14B](/models/qwen3-14b.md) | Q4_K_M | 10.9 | fit | 55.7 [estimate] |
| [gpt-oss-20b](/models/gpt-oss-20b.md) | MXFP4 | 12.9 | fit | 251.6 [estimate] |
| [Mistral Small 3.2 24B](/models/mistral-small-3.2-24b.md) | Q4_K_M | 16.2 | fit | 36.8 [estimate] |
| [Gemma 3 27B IT](/models/gemma-3-27b.md) | Q4_K_M | 18.2 | fit | 32.6 [estimate] |
| [Qwen3 30B-A3B](/models/qwen3-30b-a3b.md) | Q4_K_M | 19.9 | fit | 204.9 [estimate] |
| [Qwen3 32B](/models/qwen3-32b.md) | Q4_K_M | 22.5 | tight | 26.3 [estimate] |
| [Llama 3.3 70B Instruct](/models/llama-3.3-70b.md) | Q4_K_M | 45.8 | offload | 1.4 [estimate] |
| [gpt-oss-120b](/models/gpt-oss-120b.md) | MXFP4 | 64.3 | offload | 17.3 [estimate] |