Compare · Models · llama.cpp

Gemma 3 27B vs GLM 5.3 Flash for Radeon

Gemma 3 27B is 27.4B, 128K tokens on the official card. GLM 5.3 Flash is 320B total / 18B active, 1,048,576 tokens. Q6 / Q5 / Q4 file sizes come from the listed GGUFs on each /models page, or a calculator estimate when a listing is missing. Q3 stays off. Q4 fits Gemma 3 27B on one R9700 at 4K and spills on GLM 5.3 Flash. Neither model was measured on the local desktop. Fit numbers are calculator estimates. llama.cpp first.

How we test — Measured cells cite official amdai-bench/run@1 cards. Estimates stay labeled. Q3 is estimate-only. Q1/Q2 stay off.

Sizes, active params, context

Cited, not measured on the local desktop

Gemma 3 27B

16–24 GB dense — cited, not measured

Parameters
27.4B
27.4B total · 27.4B active
Cited context
128K tokens
Official card context claim: 128K tokens.
Q6 / Q5 / Q4 files
  • Q6_K · 22.17 GB
  • Q5_K_M · 19.27 GB
  • Q4_K_M · 16.55 GB
Official card: 128K input context for the 27B size. Hybrid sliding window — 10 global layers, 52 cache a 1024-token window.

Full model page · Official card · GGUF · Can an R9700 run it? · Open in calculator

Cited, not measured on the local desktop

GLM 5.3 Flash

320B / 18B MoE — cited, not measured

Parameters
320B total / 18B active
320B total · 18B active
Cited context
1,048,576 tokens
Official card context claim: 1,048,576 tokens.
Q6 / Q5 / Q4 files
  • UD-Q6_K_XL · 291.83 GB
  • UD-Q5_K_XL · 240.31 GB
  • UD-Q4_K_XL · 199.71 GB
Official card: 320B total / 18B active. 11 of 45 layers are sparse attention (MLA, kv_lora_rank 512, K-only cache) and grow the context store plus an 11×384 indexer (raw + 2 pooled rows). 34 linear (KDA) layers hold a fixed state. max_position_embeddings 1,048,576. 288 routed experts + 1 shared, 8 per token. Consumer cards offload; both RAM lines are estimates until measured.

Full model page · Official card · GGUF · Can an R9700 run it? · Open in calculator

Fit on common cards

Existing GPU sheets only: R9700, RX 7900 XTX, RX 7900 XT, RX 7800 XT. llama.cpp, f16 KV. Totals are calculator estimates. Speed is measured only on R9700 35B cells that already exist; everything else stays not measured yet. This table lists Q6 / Q5 / Q4. Q3 stays on the calculator as a labeled estimate.

R9700 32 GB

Gemma 3 27B vs GLM 5.3 Flash on R9700
Quant Kind Gemma 3 27B 4KGemma 3 27B 32KGemma 3 27B 125K GLM 5.3 Flash 4KGLM 5.3 Flash 32KGLM 5.3 Flash 256K
Q6

Q6_K / UD-Q6_K_XL

Estimate

22.3 GB

Fits in VRAM (estimate)

not measured yet

No official local desktop cell for this model × card. Not measured yet.

24.5 GB

Fits in VRAM (estimate)

not measured yet

No official local desktop cell for this model × card. Not measured yet.

31.9 GB

Tight — little desktop headroom

not measured yet

No official local desktop cell for this model × card. Not measured yet.

276.5 GB

Does not fit this GPU

not measured yet

No official local desktop cell for this model × card. Not measured yet.

277.0 GB

Does not fit this GPU

not measured yet

No official local desktop cell for this model × card. Not measured yet.

281.3 GB

Does not fit this GPU

not measured yet

No official local desktop cell for this model × card. Not measured yet.

Q5

Q5_K_M / UD-Q5_K_XL

Estimate

19.6 GB

Fits in VRAM (estimate)

not measured yet

No official local desktop cell for this model × card. Not measured yet.

21.8 GB

Fits in VRAM (estimate)

not measured yet

No official local desktop cell for this model × card. Not measured yet.

29.1 GB

Fits in VRAM (estimate)

not measured yet

No official local desktop cell for this model × card. Not measured yet.

227.8 GB

Does not fit this GPU

not measured yet

No official local desktop cell for this model × card. Not measured yet.

228.3 GB

Does not fit this GPU

not measured yet

No official local desktop cell for this model × card. Not measured yet.

232.6 GB

Does not fit this GPU

not measured yet

No official local desktop cell for this model × card. Not measured yet.

Q4

Q4_K_M / UD-Q4_K_XL

Estimate

17.0 GB

Fits in VRAM (estimate)

not measured yet

No official local desktop cell for this model × card. Not measured yet.

19.2 GB

Fits in VRAM (estimate)

not measured yet

No official local desktop cell for this model × card. Not measured yet.

26.6 GB

Fits in VRAM (estimate)

not measured yet

No official local desktop cell for this model × card. Not measured yet.

189.4 GB

Does not fit this GPU

not measured yet

No official local desktop cell for this model × card. Not measured yet.

189.9 GB

Does not fit this GPU

not measured yet

No official local desktop cell for this model × card. Not measured yet.

194.2 GB

Does not fit this GPU

not measured yet

No official local desktop cell for this model × card. Not measured yet.

RX 7900 XTX 24 GB

Gemma 3 27B vs GLM 5.3 Flash on RX 7900 XTX
Quant Kind Gemma 3 27B 4KGemma 3 27B 32KGemma 3 27B 125K GLM 5.3 Flash 4KGLM 5.3 Flash 32KGLM 5.3 Flash 256K
Q6

Q6_K / UD-Q6_K_XL

Estimate

22.3 GB

Fits in VRAM (estimate)

not measured yet

No official local desktop cell for this model × card. Not measured yet.

24.5 GB

Does not fit this GPU

not measured yet

No official local desktop cell for this model × card. Not measured yet.

31.9 GB

Does not fit this GPU

not measured yet

No official local desktop cell for this model × card. Not measured yet.

276.5 GB

Does not fit this GPU

not measured yet

No official local desktop cell for this model × card. Not measured yet.

277.0 GB

Does not fit this GPU

not measured yet

No official local desktop cell for this model × card. Not measured yet.

281.3 GB

Does not fit this GPU

not measured yet

No official local desktop cell for this model × card. Not measured yet.

Q5

Q5_K_M / UD-Q5_K_XL

Estimate

19.6 GB

Fits in VRAM (estimate)

not measured yet

No official local desktop cell for this model × card. Not measured yet.

21.8 GB

Fits in VRAM (estimate)

not measured yet

No official local desktop cell for this model × card. Not measured yet.

29.1 GB

Does not fit this GPU

not measured yet

No official local desktop cell for this model × card. Not measured yet.

227.8 GB

Does not fit this GPU

not measured yet

No official local desktop cell for this model × card. Not measured yet.

228.3 GB

Does not fit this GPU

not measured yet

No official local desktop cell for this model × card. Not measured yet.

232.6 GB

Does not fit this GPU

not measured yet

No official local desktop cell for this model × card. Not measured yet.

Q4

Q4_K_M / UD-Q4_K_XL

Estimate

17.0 GB

Fits in VRAM (estimate)

not measured yet

No official local desktop cell for this model × card. Not measured yet.

19.2 GB

Fits in VRAM (estimate)

not measured yet

No official local desktop cell for this model × card. Not measured yet.

26.6 GB

Does not fit this GPU

not measured yet

No official local desktop cell for this model × card. Not measured yet.

189.4 GB

Does not fit this GPU

not measured yet

No official local desktop cell for this model × card. Not measured yet.

189.9 GB

Does not fit this GPU

not measured yet

No official local desktop cell for this model × card. Not measured yet.

194.2 GB

Does not fit this GPU

not measured yet

No official local desktop cell for this model × card. Not measured yet.

RX 7900 XT 20 GB

Gemma 3 27B vs GLM 5.3 Flash on RX 7900 XT
Quant Kind Gemma 3 27B 4KGemma 3 27B 32KGemma 3 27B 125K GLM 5.3 Flash 4KGLM 5.3 Flash 32KGLM 5.3 Flash 256K
Q6

Q6_K / UD-Q6_K_XL

Estimate

22.3 GB

Does not fit this GPU

not measured yet

No official local desktop cell for this model × card. Not measured yet.

24.5 GB

Does not fit this GPU

not measured yet

No official local desktop cell for this model × card. Not measured yet.

31.9 GB

Does not fit this GPU

not measured yet

No official local desktop cell for this model × card. Not measured yet.

276.5 GB

Does not fit this GPU

not measured yet

No official local desktop cell for this model × card. Not measured yet.

277.0 GB

Does not fit this GPU

not measured yet

No official local desktop cell for this model × card. Not measured yet.

281.3 GB

Does not fit this GPU

not measured yet

No official local desktop cell for this model × card. Not measured yet.

Q5

Q5_K_M / UD-Q5_K_XL

Estimate

19.6 GB

Tight — little desktop headroom

not measured yet

No official local desktop cell for this model × card. Not measured yet.

21.8 GB

Does not fit this GPU

not measured yet

No official local desktop cell for this model × card. Not measured yet.

29.1 GB

Does not fit this GPU

not measured yet

No official local desktop cell for this model × card. Not measured yet.

227.8 GB

Does not fit this GPU

not measured yet

No official local desktop cell for this model × card. Not measured yet.

228.3 GB

Does not fit this GPU

not measured yet

No official local desktop cell for this model × card. Not measured yet.

232.6 GB

Does not fit this GPU

not measured yet

No official local desktop cell for this model × card. Not measured yet.

Q4

Q4_K_M / UD-Q4_K_XL

Estimate

17.0 GB

Fits in VRAM (estimate)

not measured yet

No official local desktop cell for this model × card. Not measured yet.

19.2 GB

Tight — little desktop headroom

not measured yet

No official local desktop cell for this model × card. Not measured yet.

26.6 GB

Does not fit this GPU

not measured yet

No official local desktop cell for this model × card. Not measured yet.

189.4 GB

Does not fit this GPU

not measured yet

No official local desktop cell for this model × card. Not measured yet.

189.9 GB

Does not fit this GPU

not measured yet

No official local desktop cell for this model × card. Not measured yet.

194.2 GB

Does not fit this GPU

not measured yet

No official local desktop cell for this model × card. Not measured yet.

RX 7800 XT 16 GB

Gemma 3 27B vs GLM 5.3 Flash on RX 7800 XT
Quant Kind Gemma 3 27B 4KGemma 3 27B 32KGemma 3 27B 125K GLM 5.3 Flash 4KGLM 5.3 Flash 32KGLM 5.3 Flash 256K
Q6

Q6_K / UD-Q6_K_XL

Estimate

22.3 GB

Does not fit this GPU

not measured yet

No official local desktop cell for this model × card. Not measured yet.

24.5 GB

Does not fit this GPU

not measured yet

No official local desktop cell for this model × card. Not measured yet.

31.9 GB

Does not fit this GPU

not measured yet

No official local desktop cell for this model × card. Not measured yet.

276.5 GB

Does not fit this GPU

not measured yet

No official local desktop cell for this model × card. Not measured yet.

277.0 GB

Does not fit this GPU

not measured yet

No official local desktop cell for this model × card. Not measured yet.

281.3 GB

Does not fit this GPU

not measured yet

No official local desktop cell for this model × card. Not measured yet.

Q5

Q5_K_M / UD-Q5_K_XL

Estimate

19.6 GB

Does not fit this GPU

not measured yet

No official local desktop cell for this model × card. Not measured yet.

21.8 GB

Does not fit this GPU

not measured yet

No official local desktop cell for this model × card. Not measured yet.

29.1 GB

Does not fit this GPU

not measured yet

No official local desktop cell for this model × card. Not measured yet.

227.8 GB

Does not fit this GPU

not measured yet

No official local desktop cell for this model × card. Not measured yet.

228.3 GB

Does not fit this GPU

not measured yet

No official local desktop cell for this model × card. Not measured yet.

232.6 GB

Does not fit this GPU

not measured yet

No official local desktop cell for this model × card. Not measured yet.

Q4

Q4_K_M / UD-Q4_K_XL

Estimate

17.0 GB

Does not fit this GPU

not measured yet

No official local desktop cell for this model × card. Not measured yet.

19.2 GB

Does not fit this GPU

not measured yet

No official local desktop cell for this model × card. Not measured yet.

26.6 GB

Does not fit this GPU

not measured yet

No official local desktop cell for this model × card. Not measured yet.

189.4 GB

Does not fit this GPU

not measured yet

No official local desktop cell for this model × card. Not measured yet.

189.9 GB

Does not fit this GPU

not measured yet

No official local desktop cell for this model × card. Not measured yet.

194.2 GB

Does not fit this GPU

not measured yet

No official local desktop cell for this model × card. Not measured yet.

System RAM

When a shopper file does not fit, the spill is system RAM. Shopper lines stay Minimum system RAM needed and RAM needed at full context on one R9700 at the model’s shopper quant, 32K selected and the official-card full context. Estimate unless a measured R9700 card overlays the kit.

Gemma 3 27B

Q4_K_M at 4K or 32K does not fit every listed card on the calculator.

Minimum system RAM needed

32 GB kit

Estimate · 24.7 GiB raw · 32K context

RAM needed at full context

32 GB kit

Estimate · 24.6 GiB raw · 8K full context

Estimate. Offloaded weights + KV in RAM + runtime + OS/desktop + headroom, rounded up to a 32 GB DIMM kit.

Breakdown
Offloaded weights ·
0 GiB
Separable embed in RAM ·
0 GiB
KV in RAM at selected ·
0 GiB
KV in RAM at full context ·
0 GiB
Runtime / backend ·
0.68 GiB
OS / desktop baseline ·
8.00 GiB
Headroom ·
16.0 GiB

GLM 5.3 Flash

UD-Q4_K_XL at 4K or 32K does not fit every listed card on the calculator.

Minimum system RAM needed

192 GB kit

Estimate · 181.9 GiB raw · 32K context

RAM needed at full context

192 GB kit

Estimate · 181.5 GiB raw · 8K full context

Estimate. Offloaded weights + KV in RAM + runtime + OS/desktop + headroom, rounded up to a 192 GB DIMM kit. The local desktop ships 256 GB DDR5; measured host-RAM usage is not recorded until Local AI numbers arrive.

Breakdown
Offloaded weights ·
154.0 GiB
Separable embed in RAM ·
0 GiB
KV in RAM at selected ·
0.74 GiB
KV in RAM at full context ·
0.29 GiB
Runtime / backend ·
3.20 GiB
OS / desktop baseline ·
8.00 GiB
Headroom ·
16.0 GiB

llama.cpp first

Commands and recipes already on the model pages. Cited-model shapes are not measured local desktop recipes. House Flash keeps its existing Strata recipe plus the llama.cpp backend page.

Gemma 3 27B

  • Gemma 3 27B llama.cpp Q4_K_M llama.cpp

    # Shape only — not a measured local desktop recipe.
    ./build-vk/bin/llama-server \
      -m ~/models/gemma-3-27b-it-Q4_K_M.gguf \
      -ngl 99 \
      -c 8192 \
      --cache-type-k f16 \
      --cache-type-v f16 \
      --host 127.0.0.1 --port 8080

GLM 5.3 Flash

  • GLM 5.3 Flash llama.cpp UD-Q4_K_XL llama.cpp

    # Shape only — not a measured local desktop recipe.
    ./build-vk/bin/llama-server \
      -m ~/models/UD-Q4_K_XL/GLM-5.3-Flash-UD-Q4_K_XL-00001-of-00006.gguf \
      -ngl 99 \
      -c 8192 \
      --cache-type-k f16 \
      --cache-type-v f16 \
      --host 127.0.0.1 --port 8080

Plain-English pick

  • Chat

    Gemma 3 27B holds Q4 on one R9700 at 4K. GLM 5.3 Flash needs RAM offload.

  • Coding agents

    Keep llama.cpp first. The side that holds Q4 in VRAM (Gemma 3 27B) is the less painful coding-agent host. House Flash / agent pins stay R9700 / local desktop — they do not transfer. No invented agent speed.

  • Long context

    Cited context is 128K tokens vs 1,048,576 tokens. KV grows with context. Q4 at 32K fits Gemma 3 27B and spills on GLM 5.3 Flash on one R9700. Read /long-context for measured R9700 35B 128K–262K cards. We do not invent a 128K decode figure here.

  • Size and VRAM

    Gemma 3 27B files: Q6_K 22.17 GB, Q5_K_M 19.27 GB, Q4_K_M 16.55 GB. GLM 5.3 Flash files: UD-Q6_K_XL 291.83 GB, UD-Q5_K_XL 240.31 GB, UD-Q4_K_XL 199.71 GB. Fit on the four existing GPU sheets below. Empty speed cells stay not measured yet.

Stack Lab

Open in Stack Lab only where the handoff is already allowlisted. Doctor is always here. House daily / Flash / agent profiles appear only when that house model is on the pair.

Open in Stack Lab

If the app is not installed, packages are not published yet. There is no public download URL. Read the Stack Lab install notes . If Stack Lab is already running on this machine, open the local app .

Pages: Gemma 3 27B · GLM 5.3 Flash · All model compares · Compare GPUs · Compare 35B · Benchmarks · Stack Lab · Long context · Calculator

Next steps

Real pages on this site. We do not invent a next click.

Last updated

esc

Type to filter. Enter opens the first match. Full-text search lives on /search.