---

license: mit

base_model: microsoft/Phi-4-reasoning-vision-15B

base_model_relation: quantized

tags: [gguf, llama.cpp, rocm, gfx1151, strix-halo, amd, ryzen-ai-max-395, ai-max-395, rocmfpx, phi, vision, multimodal, reasoning]

language: [en]

pipeline_tag: text-generation

---

> ### ⚠️ STOCK llama.cpp WILL NOT LOAD THIS MODEL

> ⚠️ -fa off is required — flash attention breaks the vision path on gfx1151.

>

> 🚀 24.62 tok/s — +9.4% faster and 0.5 GiB smaller than Q4_K_M, ranges disjoint.

Phi-4-Reasoning-Vision-15B — ROCmFP4 (tier 102 COHERENT) GGUF

A 4-bit ROCmFP4 quantization for AMD gfx1151 (Ryzen AI MAX+ 395 / Strix Halo), quantized

from F16 GGUF (27.96 GiB) — a lossless source, not a requantization of a lower-bit build.

| | |

|---|---|

| File | Phi-4-reasoning-vision-15B-Q4_0_ROCMFP4_COHERENT.gguf |

| Size | 8.0 GiB |

| BPW | 4.65 |

| ftype | Q4_0_ROCMFP4_COHERENT (102) |

⛔ Requires a llama.cpp with the ROCmFP4 quant types

Q4_0_ROCMFP4_COHERENT (ftype 102) exists only in

charlie12345/ROCmFPX, not upstream llama.cpp.

Ignore the auto-generated "Use this model" commands above.

---

All quant variants

Three builds, all measured in one session on one box with one binary (Ryzen AI MAX+ 395,

gfx1151, ROCm 7.2.4, ROCmFPX-2809dc5, -fa off) — so these rows are directly comparable.

Median of 3, warm-up discarded, otherwise-idle box.

| variant | ftype | size | bpw | decode (median) | range | repo |

|---|---|---|---|---|---|---|

| 4-bit COHERENT | 102 | 7.93 GiB | 4.65 | 24.91 | 24.88 – 24.91 | link |

| 8-bit AGENT | 115 | 14.34 GiB | 8.40 | 14.08 | 14.06 – 14.09 | link |

| 8-bit plain | 111 | 14.11 GiB | 8.27 | 14.15 | 14.13 – 14.20 | link |

⚠️ The 4-bit build is ~1.7× faster and 43% smaller. The 8-bit builds exist for accuracy

headroom, not throughput. The two 8-bit builds are within noise of each other (14.08 vs

14.15, ranges touching) — this model has no MTP draft head, and AGENT's benefit shows up in

draft acceptance, so there is nothing here for it to win.

⚠️ -fa off is mandatory — flash attention breaks the vision path on gfx1151.

The BF16 mmproj (862 MiB) ships in every one of these repos and is required for vision.

Measured

Ryzen AI MAX+ 395 (gfx1151, 128 GB unified, ROCm 7.2.4). **Median of 3+, warm-up discarded,

otherwise-idle box.** Correctness at the model's official sampling.

| build | size | decode (median) | range |

|---|---|---|---|

| this build | 8.0 GiB | 24.62 | [23.42 – 25.37] |

| Q4_K_M | 8.5 GiB | 22.50 | [22.25 – 22.50] |

+9.4%, ranges disjoint. Vision verified: an 8×8 red PNG is correctly described as Red.

Correctness: 17×23 ⇒ ✅ 391 · capital of Japan ⇒ ✅ Tokyo · days in 2024 ⇒ ✅ 366

Per-tensor types (audited in the finished file)

output.weight Q6_K · token_embd Q6_K · norms F32 · bulk TYPE_100 (160) · vision tower kept at BF16 (mmproj, 823 MiB — included)

Serve with -fa off and the bundled BF16 mmproj.

---

What was NOT measured

• No perplexity run, and no quality A/B against the baseline or the source. The checks

above are memorized-fact prompts — necessary but not sufficient; a damaged model can pass them.

• No long-context testing.
• No tool-calling evaluation.

Base model licence inherited; credit for the model goes to its authors.