R9700 vs RX 7900 GRE for local AI

VRAM, bandwidth, gfx

Which 35B quants fit

Qwen3.6-35B-A3B on llama.cpp, f16 KV. Totals are calculator estimates. Speed is measured only on the R9700 cells that already exist; everything else stays not measured. UD-Q3_K_XL is always a labeled estimate.

Weights file Estimate: total_params × 6.56 bpw (nominal GGUF Q6_K; UD-Q6_K_XL uses the same bucket) / 8. Active params do not shrink the file.. Same GGUF on both cards.

System RAM

When a house 35B Q4 file does not fit, the spill is system RAM. Shopper lines stay Minimum system RAM needed and RAM needed at full context for Qwen3.6-35B-A3B at llama.cpp, Q4_K_M, 32K selected and the model’s full context. Estimate unless a measured R9700 card overlays the kit. If both sides fit, those labels still name the calculator lines if you raise context or a larger quant.

ROCm and Vulkan

1. Note 1

   llama.cpp first

   Vulkan llama.cpp is the consumer default. HIP when rocminfo lists the gfx on this page. Ollama and vLLM are not the house path on these sheets.

2. Note 2

   Different gfx family packages

   R9700 needs the gfx1201 family package (First-class on ROCm 10.0 / 10.1. Our test-rig daily driver.) RX 7900 GRE needs gfx1100 (gfx1100, 16 GB.) Do not mix those pins.

3. Note 3

   Cited support rows

   R9700: official — First-class on ROCm 10.0 / 10.1. Our test-rig daily driver. RX 7900 GRE: official — gfx1100, 16 GB.

Cited rows: ROCm 10.1.0 compatibility matrix and Linux system requirements. More gfx notes on /rocm/gfx.

Plain-English pick

• Chat

  R9700 holds house 35B Q4 at 4K in the calculator VRAM pool. RX 7900 GRE needs RAM offload or a listed UD-Q3_K_XL file — that Q3 is a labeled estimate (Hugging Face listing for Qwen3.6-35B-A3B-UD-Q3_K_XL.gguf (16.8 GB)).

• Coding agents

  Coding agents want the file in VRAM and a stable llama.cpp server. The house daily-driver and agent recipes are measured on the R9700. RX 7900 GRE is 16 GB — size Q4 on the calculator before you assume that recipe applies. We do not invent an agent tok/s for RX 7900 GRE.

• Long context

  Long context is a KV-cache problem. Measured 128K–262K 35B llama.cpp cards live on the R9700 page and /long-context. RX 7900 GRE has less headroom — Q4 at 32K is a calculator spill on that card. Empty speed cells stay not measured.

• Image gen

  Image gen on AMDai is ComfyUI / HIP PyTorch, not llama.cpp. R9700 is the 32 GB discrete class; RX 7900 GRE is the 16 GB discrete class. Seconds-per-image stay not yet measured by AMDai. Start smaller (SD 1.5) on 8 / 12 / 16 GB class cards. Flux-class is tighter than SDXL. Size the checkpoint on the image-gen guide — do not copy a tok/s from the LLM table.

Check price

Affiliate hops only. No live or invented street price. Same disclosure as every other sheet.