NVIDIA GeForce RTX 4090: which models fit?

RTX 4090 RTX 5090 RTX 3090 RTX 4060 Ti (16 GB)RTX 5070 Ti Radeon RX 7900 XTX M4 Max 128 GB M3 Ultra 512 GB DGX Spark (128 GB)All devices

NVIDIA · GPU

NVIDIA GeForce RTX 4090

Memory24 GB GDDR6XBandwidth1,008 GB/s vendor specPlanned for models24.0 GiB all VRAM usableFilesGGUF llama.cpp format

Editorial pick · 24 GB GPUsQwen3.8-27B Q5_K_M All picks

Budget = file + KV cache + 1 GiB runtime + 0.5 GiB spare. Macs: 75% of memory. Details

34 models fit RTX 4090 · 8K context

Ranked by Artificial Analysis Intelligence Index, unranked models last. Arena is shown separately. How ranking works

Intelligence

Artificial Analysis Intelligence Index · top 10 models that fit RTX 4090 · higher is better

AA estimate

Generation speed

Tokens per second on RTX 4090, best file for each of the same models · higher is better

MeasuredEstimated (~)

WeightsKV cacheRuntimeFreeMeasuredlinks to its sourceEstimatedshown with ~Unknownnever zero

Qwen3.8-27B

Qwen/Qwen3.8-27B

2026-08-1427B dense

AA index33.7rank #13

Arena1,441

Best file & memory

Q5_K_M19.5 GiB fileFits · conditional

21.0 of 24.0 GiB+8K ctx · max ≈48K

Prompt~2,965 t/sEstimated

Generate~30.3 t/sEstimated

First token · 4K prompt~1.4 sEstimated

Memory budget

Weights file	19.49 GiB	bartowski/Qwen3.8-27B-GGUF HF files
KV cache · 8K	0.50 GiB	FP16 cache from config.json Estimated
Runtime allowance	1.00 GiB	planning assumption Estimated
Recurrent state	—	Unknown
Accounted	20.99 GiB	of 24.0 GiB usable (100% of 24 GB) · 0.5 GiB spare margin required

Max context in free memory: ≈48K tokens (native 256K).

Other GGUF files

Q4_K_M 16.2 GiB · fits · files
Q8_0 27.1 GiB · too big · files

Where the speeds come from

Prompt estimated from median of 20 measured prefill points for this device (llama.cpp CUDA), rescaled to this model.
Generate estimated from median of 24 measured decode points for this device (llama.cpp CUDA), rescaled to this model.
First token 4K-token prompt ÷ estimated prefill speed. Excludes model load and queueing.

Caveats

Fixed recurrent/convolution state is unknown and not included.
Text-only budget. A downloadable file does not prove your runtime supports this architecture.

AA 33.7 Qwen3.8 27B (Xhigh) · Arena 1,441 · Model page

G9v3-39A5B (preview)

ai9stars/G9v3-39A5B

2026-08 (day unknown)39B · 5B active

AA index21.8 AA estimaterank #29

Arena—

Best file & memory

Q4_K_M21.9 GiB fileFits

23.2 of 24.0 GiB8K ctx · max ≈15K

Prompt~6,633 t/sEstimated

Generate~160 t/sEstimated

First token · 4K prompt~618 msEstimated

Memory budget

Weights file	21.94 GiB	linuxid10t/G9v3-39A5B-GGUF HF files
KV cache · 8K	0.30 GiB	FP16 cache from config.json Estimated
Runtime allowance	1.00 GiB	planning assumption Estimated
Accounted	23.24 GiB	of 24.0 GiB usable (100% of 24 GB) · 0.5 GiB spare margin required

Max context in free memory: ≈15K tokens (native 128K).

Other GGUF files

Q8_0 38.6 GiB · too big · files

Where the speeds come from

Prompt estimated from median of 4 measured prefill points for this device (llama.cpp CUDA), rescaled to this model.
Generate estimated from median of 10 measured decode points for this device (llama.cpp CUDA), rescaled to this model.
First token 4K-token prompt ÷ estimated prefill speed. Excludes model load and queueing.

Caveats

Text-only budget. A downloadable file does not prove your runtime supports this architecture.

AA 21.8 AA estimate G9v3-39A5B · Arena unknown · Model page

K2-Horizon-7B

IFM/K2-Horizon-7B

2026-09-03size unknown

AA index20.6rank #33

Arena—

Best file & memory

Q8_08.9 GiB fileFits

11.0 of 24.0 GiB8K ctx · max ≈97K

Prompt—Unknown

Generate~66.2 t/sEstimated

First token · 4K prompt—Unknown

Memory budget

Weights file	8.92 GiB	IFM/K2-Horizon-7B-GGUF HF files
KV cache · 8K	1.13 GiB	FP16 cache from config.json Estimated
Runtime allowance	1.00 GiB	planning assumption Estimated
Accounted	11.04 GiB	of 24.0 GiB usable (100% of 24 GB) · 0.5 GiB spare margin required

Max context in free memory: ≈97K tokens (native 512K).

Other GGUF files

Q4_K_M 5.2 GiB · fits · files
Q5_K_M 6.0 GiB · fits · files
FP16 16.8 GiB · fits · files

Where the speeds come from

Prompt unknown: no measured report, and no comparable vendor FP16 figure or device calibration.
Generate estimated from median of 24 measured decode points for this device (llama.cpp CUDA), rescaled to this model.
First token unknown: needs a prefill value.

Caveats

Text-only budget. A downloadable file does not prove your runtime supports this architecture.

AA 20.6 K2 Horizon 7B · Arena unknown · Model page

Qwen3.6-35B-A3B

Qwen/Qwen3.6-35B-A3B

2026-04-1635B · 3B active

AA index18.2rank #36

Arena—

Best file & memory

Q4_K_M19.0 GiB fileFits · conditional

20.2 of 24.0 GiB+8K ctx · max ≈178K

Prompt~11,055 t/sEstimated

Generate~277 t/sEstimated

First token · 4K prompt~371 msEstimated

Memory budget

Weights file	19.02 GiB	ggml-org/Qwen3.6-35B-A3B-GGUF HF files
KV cache · 8K	0.16 GiB	FP16 cache from config.json Estimated
Runtime allowance	1.00 GiB	planning assumption Estimated
Recurrent state	—	Unknown
Accounted	20.17 GiB	of 24.0 GiB usable (100% of 24 GB) · 0.5 GiB spare margin required

Max context in free memory: ≈178K tokens (native 256K).

Other GGUF files

Q5_K_M 23.0 GiB · too big · files
Q8_0 34.4 GiB · too big · files

Where the speeds come from

Prompt estimated from median of 4 measured prefill points for this device (llama.cpp CUDA), rescaled to this model.
Generate estimated from median of 10 measured decode points for this device (llama.cpp CUDA), rescaled to this model.
First token 4K-token prompt ÷ estimated prefill speed. Excludes model load and queueing.

Caveats

Fixed recurrent/convolution state is unknown and not included.
Text-only budget. A downloadable file does not prove your runtime supports this architecture.

AA 18.2 Qwen3.6 35B A3B (Reasoning) · Arena unknown · Model page

Muse Glimmer 30B

meta-models/Muse-Glimmer-30B

2026-08-1030B dense

AA index17.5rank #38

Arena1,389

Best file & memory

Q5_K_M18.7 GiB fileFits

19.9 of 24.0 GiB8K ctx · max ≈128K

Prompt~2,704 t/sEstimated

Generate~31.5 t/sEstimated

First token · 4K prompt~1.5 sEstimated

Memory budget

Weights file	18.72 GiB	bartowski/Muse-Glimmer-30B-GGUF HF files
KV cache · 8K	0.18 GiB	FP16 cache from config.json Estimated
Runtime allowance	1.00 GiB	planning assumption Estimated
Accounted	19.90 GiB	of 24.0 GiB usable (100% of 24 GB) · 0.5 GiB spare margin required

Max context in free memory: ≈128K tokens (native 128K).

Other GGUF files

Q4_K_M 15.6 GiB · fits · files
Q8_0 27.6 GiB · too big · files

Where the speeds come from

Prompt estimated from median of 20 measured prefill points for this device (llama.cpp CUDA), rescaled to this model.
Generate estimated from median of 24 measured decode points for this device (llama.cpp CUDA), rescaled to this model.
First token 4K-token prompt ÷ estimated prefill speed. Excludes model load and queueing.

Caveats

Text-only budget. A downloadable file does not prove your runtime supports this architecture.

AA 17.5 Muse Glimmer (High) · Arena 1,389 · Model page

K2-Horizon-3.7B

IFM/K2-Horizon-3.7B

2026-09-03size unknown

AA index15.6rank #41

Arena—

Best file & memory

Q8_05.0 GiB fileFits

7.1 of 24.0 GiB8K ctx · max ≈124K

Prompt—Unknown

Generate~118 t/sEstimated

First token · 4K prompt—Unknown

Memory budget

Weights file	5.02 GiB	IFM/K2-Horizon-3.7B-GGUF HF files
KV cache · 8K	1.13 GiB	FP16 cache from config.json Estimated
Runtime allowance	1.00 GiB	planning assumption Estimated
Accounted	7.14 GiB	of 24.0 GiB usable (100% of 24 GB) · 0.5 GiB spare margin required

Max context in free memory: ≈124K tokens (native 512K).

Other GGUF files

Q4_K_M 2.9 GiB · fits · files
Q5_K_M 3.4 GiB · fits · files
FP16 9.4 GiB · fits · files

Where the speeds come from

Prompt unknown: no measured report, and no comparable vendor FP16 figure or device calibration.
Generate estimated from median of 24 measured decode points for this device (llama.cpp CUDA), rescaled to this model.
First token unknown: needs a prefill value.

Caveats

Text-only budget. A downloadable file does not prove your runtime supports this architecture.

AA 15.6 K2 Horizon 3.7B · Arena unknown · Model page

Qwen3.5-9B

Qwen/Qwen3.5-9B

2026-03-029B dense

AA index13.3 AA estimaterank #45

Arena—

Best file & memory

Q8_08.9 GiB fileFits · conditional

10.1 of 24.0 GiB+8K ctx · max ≈256K

Prompt~8,894 t/sEstimated

Generate~66.5 t/sEstimated

First token · 4K prompt~461 msEstimated

Memory budget

Weights file	8.87 GiB	unsloth/Qwen3.5-9B-GGUF HF files
KV cache · 8K	0.25 GiB	FP16 cache from config.json Estimated
Runtime allowance	1.00 GiB	planning assumption Estimated
Recurrent state	—	Unknown
Accounted	10.12 GiB	of 24.0 GiB usable (100% of 24 GB) · 0.5 GiB spare margin required

Max context in free memory: ≈256K tokens (native 256K).

Other GGUF files

Q4_K_M 5.3 GiB · fits · files
Q5_K_M 6.1 GiB · fits · files

Where the speeds come from

Prompt estimated from median of 20 measured prefill points for this device (llama.cpp CUDA), rescaled to this model.
Generate estimated from median of 24 measured decode points for this device (llama.cpp CUDA), rescaled to this model.
First token 4K-token prompt ÷ estimated prefill speed. Excludes model load and queueing.

Caveats

Fixed recurrent/convolution state is unknown and not included.
Text-only budget. A downloadable file does not prove your runtime supports this architecture.

AA 13.3 AA estimate Qwen3.5 9B (Non-reasoning) · Arena unknown · Model page

EXAONE 4.5 33B

LGAI-EXAONE/EXAONE-4.5-33B

2026-04-0933B dense

AA index13.2 AA estimaterank #46

Arena—

Best file & memory

Q4_K_M18.7 GiB fileFits

20.9 of 24.0 GiB8K ctx · max ≈49K

Prompt~2,426 t/sEstimated

Generate~31.6 t/sEstimated

First token · 4K prompt~1.7 sEstimated

Memory budget

Weights file	18.67 GiB	LGAI-EXAONE/EXAONE-4.5-33B-GGUF HF files
KV cache · 8K	1.25 GiB	FP16 cache from config.json Estimated
Runtime allowance	1.00 GiB	planning assumption Estimated
Accounted	20.92 GiB	of 24.0 GiB usable (100% of 24 GB) · 0.5 GiB spare margin required

Max context in free memory: ≈49K tokens (native 256K).

Other GGUF files

Q5_K_M 21.9 GiB · tight · files
Q8_0 32.7 GiB · too big · files

Where the speeds come from

Prompt estimated from median of 20 measured prefill points for this device (llama.cpp CUDA), rescaled to this model.
Generate estimated from median of 24 measured decode points for this device (llama.cpp CUDA), rescaled to this model.
First token 4K-token prompt ÷ estimated prefill speed. Excludes model load and queueing.

Caveats

Text-only budget. A downloadable file does not prove your runtime supports this architecture.

AA 13.2 AA estimate EXAONE 4.5 33B (Reasoning) · Arena unknown · Model page

Qwen3.5-4B

Qwen/Qwen3.5-4B

2026-03-024B dense

AA index13.1 AA estimaterank #48

Arena—

Best file & memory

Q8_04.2 GiB fileFits · conditional

5.4 of 24.0 GiB+8K ctx · max ≈256K

Prompt~20,013 t/sEstimated

Generate~141 t/sEstimated

First token · 4K prompt~205 msEstimated

Memory budget

Weights file	4.17 GiB	unsloth/Qwen3.5-4B-GGUF HF files
KV cache · 8K	0.25 GiB	FP16 cache from config.json Estimated
Runtime allowance	1.00 GiB	planning assumption Estimated
Recurrent state	—	Unknown
Accounted	5.42 GiB	of 24.0 GiB usable (100% of 24 GB) · 0.5 GiB spare margin required

Max context in free memory: ≈256K tokens (native 256K).

Other GGUF files

Q4_K_M 2.6 GiB · fits · files
Q5_K_M 2.9 GiB · fits · files

Where the speeds come from

Prompt estimated from median of 20 measured prefill points for this device (llama.cpp CUDA), rescaled to this model.
Generate estimated from median of 24 measured decode points for this device (llama.cpp CUDA), rescaled to this model.
First token 4K-token prompt ÷ estimated prefill speed. Excludes model load and queueing.

Caveats

Fixed recurrent/convolution state is unknown and not included.
Text-only budget. A downloadable file does not prove your runtime supports this architecture.

AA 13.1 AA estimate Qwen3.5 4B (Reasoning) · Arena unknown · Model page

IBM Granite 4.2 30B

ibm-granite/granite-4.2-30b

2026-08-2530B dense

AA index12.7rank #50

Arena1,361

Best file & memory

Q5_K_M19.3 GiB fileFits

22.3 of 24.0 GiB8K ctx · max ≈13K

Prompt~2,668 t/sEstimated

Generate~30.5 t/sEstimated

First token · 4K prompt~1.5 sEstimated

Memory budget

Weights file	19.35 GiB	ibm-granite/granite-4.2-30b-GGUF HF files
KV cache · 8K	2.00 GiB	FP16 cache from config.json Estimated
Runtime allowance	1.00 GiB	planning assumption Estimated
Accounted	22.35 GiB	of 24.0 GiB usable (100% of 24 GB) · 0.5 GiB spare margin required

Max context in free memory: ≈13K tokens (native 128K).

Other GGUF files

Q4_K_M 16.5 GiB · fits · files
Q8_0 29.0 GiB · too big · files

Where the speeds come from

Prompt estimated from median of 20 measured prefill points for this device (llama.cpp CUDA), rescaled to this model.
Generate estimated from median of 24 measured decode points for this device (llama.cpp CUDA), rescaled to this model.
First token 4K-token prompt ÷ estimated prefill speed. Excludes model load and queueing.

Caveats

Text-only budget. A downloadable file does not prove your runtime supports this architecture.

AA 12.7 Granite 4.2 30B · Arena 1,361 · Model page

MiniCPM5-2B

openbmb/MiniCPM5-2B

2026-09-072.5B dense

AA index12.5rank #51

Arena—

Best file & memory

Q8_02.5 GiB fileFits

3.8 of 24.0 GiB8K ctx · max ≈128K

Prompt~31,807 t/sEstimated

Generate~236 t/sEstimated

First token · 4K prompt~129 msEstimated

Memory budget

Weights file	2.50 GiB	openbmb/MiniCPM5-2B-GGUF HF files
KV cache · 8K	0.33 GiB	FP16 cache from config.json Estimated
Runtime allowance	1.00 GiB	planning assumption Estimated
Accounted	3.82 GiB	of 24.0 GiB usable (100% of 24 GB) · 0.5 GiB spare margin required

Max context in free memory: ≈128K tokens (native 128K).

Other GGUF files

Q4_K_M 1.5 GiB · fits · files
Q5_K_M 1.8 GiB · fits · files
FP16 4.7 GiB · fits · files

Where the speeds come from

Prompt estimated from median of 20 measured prefill points for this device (llama.cpp CUDA), rescaled to this model.
Generate estimated from median of 24 measured decode points for this device (llama.cpp CUDA), rescaled to this model.
First token 4K-token prompt ÷ estimated prefill speed. Excludes model load and queueing.

Caveats

Text-only budget. A downloadable file does not prove your runtime supports this architecture.

AA 12.5 MiniCPM5-2B · Arena unknown · Model page

Granite 4.2 8B

ibm-granite/granite-4.2-8b

2026-08-258B dense

AA index11.1rank #56

Arena1,320

Best file & memory

Q8_08.7 GiB fileFits

11.0 of 24.0 GiB8K ctx · max ≈88K

Prompt~10,006 t/sEstimated

Generate~67.8 t/sEstimated

First token · 4K prompt~409 msEstimated

Memory budget

Weights file	8.70 GiB	ibm-granite/granite-4.2-8b-GGUF HF files
KV cache · 8K	1.25 GiB	FP16 cache from config.json Estimated
Runtime allowance	1.00 GiB	planning assumption Estimated
Accounted	10.95 GiB	of 24.0 GiB usable (100% of 24 GB) · 0.5 GiB spare margin required

Max context in free memory: ≈88K tokens (native 128K).

Other GGUF files

Q4_K_M 5.0 GiB · fits · files
Q5_K_M 5.8 GiB · fits · files
FP16 16.4 GiB · fits · files

Where the speeds come from

Prompt estimated from median of 20 measured prefill points for this device (llama.cpp CUDA), rescaled to this model.
Generate estimated from median of 24 measured decode points for this device (llama.cpp CUDA), rescaled to this model.
First token 4K-token prompt ÷ estimated prefill speed. Excludes model load and queueing.

Caveats

Text-only budget. A downloadable file does not prove your runtime supports this architecture.

AA 11.1 Granite 4.2 8B · Arena 1,320 · Model page

G9v3-3B

AI9Stars/G9v3-3B

2026-07 (day unknown)3B dense

AA index10.8 AA estimaterank #58

Arena—

Best file & memory

Q8_03.0 GiB fileFits

4.4 of 24.0 GiB8K ctx · max ≈128K

Prompt~26,683 t/sEstimated

Generate~199 t/sEstimated

First token · 4K prompt~154 msEstimated

Memory budget

Weights file	2.96 GiB	mradermacher/G9v3-3B-GGUF HF files
KV cache · 8K	0.41 GiB	FP16 cache from config.json Estimated
Runtime allowance	1.00 GiB	planning assumption Estimated
Accounted	4.37 GiB	of 24.0 GiB usable (100% of 24 GB) · 0.5 GiB spare margin required

Max context in free memory: ≈128K tokens (native 128K).

Other GGUF files

Q4_K_M 1.7 GiB · fits · files
Q5_K_M 2.0 GiB · fits · files
FP16 5.6 GiB · fits · files

Where the speeds come from

Prompt estimated from median of 20 measured prefill points for this device (llama.cpp CUDA), rescaled to this model.
Generate estimated from median of 24 measured decode points for this device (llama.cpp CUDA), rescaled to this model.
First token 4K-token prompt ÷ estimated prefill speed. Excludes model load and queueing.

Caveats

Text-only budget. A downloadable file does not prove your runtime supports this architecture.

AA 10.8 AA estimate G9v3-3B · Arena unknown · Model page

North Mini Code

CohereLabs/North-Mini-Code-1.0

2026-06-0930B · 3B active

AA index9.9rank #61

Arena—

Best file & memory

Q5_K_M20.3 GiB fileFits

21.8 of 24.0 GiB8K ctx · max ≈74K

Prompt~11,055 t/sEstimated

Generate~222 t/sEstimated

First token · 4K prompt~371 msEstimated

Memory budget

Weights file	20.34 GiB	bartowski/North-Mini-Code-1.0-GGUF HF files
KV cache · 8K	0.48 GiB	FP16 cache from config.json Estimated
Runtime allowance	1.00 GiB	planning assumption Estimated
Accounted	21.83 GiB	of 24.0 GiB usable (100% of 24 GB) · 0.5 GiB spare margin required

Max context in free memory: ≈74K tokens (native 250K).

Other GGUF files

Q4_K_M 17.5 GiB · fits · files
Q8_0 30.2 GiB · too big · files

Where the speeds come from

Prompt estimated from median of 4 measured prefill points for this device (llama.cpp CUDA), rescaled to this model.
Generate estimated from median of 10 measured decode points for this device (llama.cpp CUDA), rescaled to this model.
First token 4K-token prompt ÷ estimated prefill speed. Excludes model load and queueing.

Caveats

Text-only budget. A downloadable file does not prove your runtime supports this architecture.

AA 9.9 North Mini Code · Arena unknown · Model page

Granite 4.2 3B

ibm-granite/granite-4.2-3b

2026-08-253B dense

AA index9.1rank #64

Arena1,293

Best file & memory

Q8_03.6 GiB fileFits

5.3 of 24.0 GiB8K ctx · max ≈128K

Prompt~26,683 t/sEstimated

Generate~163 t/sEstimated

First token · 4K prompt~154 msEstimated

Memory budget

Weights file	3.63 GiB	ibm-granite/granite-4.2-3b-GGUF HF files
KV cache · 8K	0.63 GiB	FP16 cache from config.json Estimated
Runtime allowance	1.00 GiB	planning assumption Estimated
Accounted	5.25 GiB	of 24.0 GiB usable (100% of 24 GB) · 0.5 GiB spare margin required

Max context in free memory: ≈128K tokens (native 128K).

Other GGUF files

Q4_K_M 2.1 GiB · fits · files
Q5_K_M 2.4 GiB · fits · files
FP16 6.8 GiB · fits · files

Where the speeds come from

Prompt estimated from median of 20 measured prefill points for this device (llama.cpp CUDA), rescaled to this model.
Generate estimated from median of 24 measured decode points for this device (llama.cpp CUDA), rescaled to this model.
First token 4K-token prompt ÷ estimated prefill speed. Excludes model load and queueing.

Caveats

Text-only budget. A downloadable file does not prove your runtime supports this architecture.

AA 9.1 Granite 4.2 3B · Arena 1,293 · Model page

Tri-21B-Think

trillionlabs/Tri-21B-Think

2026-02 (day unknown)21B dense

AA index9.0 AA estimaterank #65

Arena—

Best file & memory

Q8_020.5 GiB fileFits

23.1 of 24.0 GiB8K ctx · max ≈10K

Prompt~3,862 t/sEstimated

Generate~28.8 t/sEstimated

First token · 4K prompt~1.1 sEstimated

Memory budget

Weights file	20.52 GiB	mykor/Tri-21B-Think-GGUF HF files
KV cache · 8K	1.56 GiB	FP16 cache from config.json Estimated
Runtime allowance	1.00 GiB	planning assumption Estimated
Accounted	23.08 GiB	of 24.0 GiB usable (100% of 24 GB) · 0.5 GiB spare margin required

Max context in free memory: ≈10K tokens (native 32K).

Other GGUF files

Q4_K_M 11.7 GiB · fits · files
Q5_K_M 13.7 GiB · fits · files
FP16 38.6 GiB · too big · files

Where the speeds come from

Prompt estimated from median of 20 measured prefill points for this device (llama.cpp CUDA), rescaled to this model.
Generate estimated from median of 24 measured decode points for this device (llama.cpp CUDA), rescaled to this model.
First token 4K-token prompt ÷ estimated prefill speed. Excludes model load and queueing.

Caveats

Text-only budget. A downloadable file does not prove your runtime supports this architecture.

AA 9.0 AA estimate Tri-21B-Think · Arena unknown · Model page

MiniCPM5-1B

openbmb/MiniCPM5-1B

2026-05-191.1B dense

AA index8.8 AA estimaterank #67

Arena—

Best file & memory

Q8_01.1 GiB fileFits

2.3 of 24.0 GiB8K ctx · max ≈128K

Prompt~74,077 t/sEstimated

Generate~549 t/sEstimated

First token · 4K prompt~55 msEstimated

Memory budget

Weights file	1.07 GiB	openbmb/MiniCPM5-1B-GGUF HF files
KV cache · 8K	0.19 GiB	FP16 cache from config.json Estimated
Runtime allowance	1.00 GiB	planning assumption Estimated
Accounted	2.26 GiB	of 24.0 GiB usable (100% of 24 GB) · 0.5 GiB spare margin required

Max context in free memory: ≈128K tokens (native 128K).

Other GGUF files

Q4_K_M 0.6 GiB · fits · files
Q5_K_M 0.7 GiB · fits · files
FP16 2.0 GiB · fits · files

Where the speeds come from

Prompt estimated from median of 20 measured prefill points for this device (llama.cpp CUDA), rescaled to this model.
Generate estimated from median of 24 measured decode points for this device (llama.cpp CUDA), rescaled to this model.
First token 4K-token prompt ÷ estimated prefill speed. Excludes model load and queueing.

Caveats

Text-only budget. A downloadable file does not prove your runtime supports this architecture.

AA 8.8 AA estimate MiniCPM5-1B (Reasoning) · Arena unknown · Model page

Nanbeige4.1-3B

Nanbeige/Nanbeige4.1-3B

2026-02 (day unknown)3B dense

AA index8.4 AA estimaterank #69

Arena—

Best file & memory

Q8_03.9 GiB fileFits

5.4 of 24.0 GiB8K ctx · max ≈256K

Prompt~26,683 t/sEstimated

Generate~151 t/sEstimated

First token · 4K prompt~154 msEstimated

Memory budget

Weights file	3.90 GiB	mradermacher/Nanbeige4.1-3B-GGUF HF files
KV cache · 8K	0.50 GiB	FP16 cache from config.json Estimated
Runtime allowance	1.00 GiB	planning assumption Estimated
Accounted	5.40 GiB	of 24.0 GiB usable (100% of 24 GB) · 0.5 GiB spare margin required

Max context in free memory: ≈256K tokens (native 256K).

Other GGUF files

Q4_K_M 2.3 GiB · fits · files
Q5_K_M 2.6 GiB · fits · files
FP16 7.3 GiB · fits · files

Where the speeds come from

Prompt estimated from median of 20 measured prefill points for this device (llama.cpp CUDA), rescaled to this model.
Generate estimated from median of 24 measured decode points for this device (llama.cpp CUDA), rescaled to this model.
First token 4K-token prompt ÷ estimated prefill speed. Excludes model load and queueing.

Caveats

Text-only budget. A downloadable file does not prove your runtime supports this architecture.

AA 8.4 AA estimate Nanbeige4.1-3B · Arena unknown · Model page

LFM2.5-2.6B

LiquidAI/LFM2.5-2.6B

2026-08-042.7B dense

AA index8.4 AA estimaterank #70

Arena—

Best file & memory

Q8_02.7 GiB fileFits · conditional

3.8 of 24.0 GiB+8K ctx · max ≈128K

Prompt~29,758 t/sEstimated

Generate~220 t/sEstimated

First token · 4K prompt~138 msEstimated

Memory budget

Weights file	2.68 GiB	LiquidAI/LFM2.5-2.6B-GGUF HF files
KV cache · 8K	0.13 GiB	FP16 cache from config.json Estimated
Runtime allowance	1.00 GiB	planning assumption Estimated
Recurrent state	—	Unknown
Accounted	3.80 GiB	of 24.0 GiB usable (100% of 24 GB) · 0.5 GiB spare margin required

Max context in free memory: ≈128K tokens (native 128K).

Other GGUF files

Q4_K_M 1.6 GiB · fits · files
Q5_K_M 1.8 GiB · fits · files
FP16 5.0 GiB · fits · files

Where the speeds come from

Prompt estimated from median of 20 measured prefill points for this device (llama.cpp CUDA), rescaled to this model.
Generate estimated from median of 24 measured decode points for this device (llama.cpp CUDA), rescaled to this model.
First token 4K-token prompt ÷ estimated prefill speed. Excludes model load and queueing.

Caveats

Fixed recurrent/convolution state is unknown and not included.
Text-only budget. A downloadable file does not prove your runtime supports this architecture.

AA 8.4 AA estimate LFM2.5-2.6B · Arena unknown · Model page

Falcon-H1R-7B

tiiuae/Falcon-H1R-7B

2026-01-057B dense

AA index7.8 AA estimaterank #71

Arena—

Best file & memory

Q8_07.5 GiB fileFits · conditional

8.9 of 24.0 GiB+8K ctx · max ≈256K

Prompt~11,436 t/sEstimated

Generate~78.5 t/sEstimated

First token · 4K prompt~358 msEstimated

Memory budget

Weights file	7.51 GiB	tiiuae/Falcon-H1R-7B-GGUF HF files
KV cache · 8K	0.34 GiB	FP16 cache from config.json Estimated
Runtime allowance	1.00 GiB	planning assumption Estimated
Recurrent state	—	Unknown
Accounted	8.86 GiB	of 24.0 GiB usable (100% of 24 GB) · 0.5 GiB spare margin required

Max context in free memory: ≈256K tokens (native 256K).

Other GGUF files

Q4_K_M 4.3 GiB · fits · files
Q5_K_M 5.0 GiB · fits · files
FP16 14.1 GiB · fits · files

Where the speeds come from

Prompt estimated from median of 20 measured prefill points for this device (llama.cpp CUDA), rescaled to this model.
Generate estimated from median of 24 measured decode points for this device (llama.cpp CUDA), rescaled to this model.
First token 4K-token prompt ÷ estimated prefill speed. Excludes model load and queueing.

Caveats

Fixed recurrent/convolution state is unknown and not included.
Text-only budget. A downloadable file does not prove your runtime supports this architecture.

AA 7.8 AA estimate Falcon-H1R-7B · Arena unknown · Model page

Step3 VL 10B

stepfun-ai/Step3-VL-10B

2026-01 (day unknown)10B dense

AA index7.7 AA estimaterank #73

Arena—

Best file & memory

Q8_08.1 GiB fileFits

10.2 of 24.0 GiB8K ctx · max ≈64K

Prompt~8,005 t/sEstimated

Generate~72.8 t/sEstimated

First token · 4K prompt~512 msEstimated

Memory budget

Weights file	8.11 GiB	JamePeng2023/Step3-VL-10B-GGUF HF files
KV cache · 8K	1.13 GiB	FP16 cache from config.json Estimated
Runtime allowance	1.00 GiB	planning assumption Estimated
Accounted	10.24 GiB	of 24.0 GiB usable (100% of 24 GB) · 0.5 GiB spare margin required

Max context in free memory: ≈64K tokens (native 64K).

Other GGUF files

Q4_K_M 4.7 GiB · fits · files
Q5_K_M 5.4 GiB · fits · files
FP16 15.3 GiB · fits · files

Where the speeds come from

Prompt estimated from median of 20 measured prefill points for this device (llama.cpp CUDA), rescaled to this model.
Generate estimated from median of 24 measured decode points for this device (llama.cpp CUDA), rescaled to this model.
First token 4K-token prompt ÷ estimated prefill speed. Excludes model load and queueing.

Caveats

Text-only budget. A downloadable file does not prove your runtime supports this architecture.

AA 7.7 AA estimate Step3 VL 10B · Arena unknown · Model page

LFM2.5-8B-A1B

LiquidAI/LFM2.5-8B-A1B

2026-05-288.3B · 1.5B active

AA index7.2 AA estimaterank #74

Arena—

Best file & memory

Q8_08.4 GiB fileFits · conditional

9.5 of 24.0 GiB+8K ctx · max ≈125K

Prompt~22,110 t/sEstimated

Generate~297 t/sEstimated

First token · 4K prompt~185 msEstimated

Memory budget

Weights file	8.39 GiB	LiquidAI/LFM2.5-8B-A1B-GGUF HF files
KV cache · 8K	0.09 GiB	FP16 cache from config.json Estimated
Runtime allowance	1.00 GiB	planning assumption Estimated
Recurrent state	—	Unknown
Accounted	9.49 GiB	of 24.0 GiB usable (100% of 24 GB) · 0.5 GiB spare margin required

Max context in free memory: ≈125K tokens (native 125K).

Other GGUF files

Q4_K_M 4.8 GiB · fits · files
Q5_K_M 5.6 GiB · fits · files
FP16 15.8 GiB · fits · files

Where the speeds come from

Prompt estimated from median of 4 measured prefill points for this device (llama.cpp CUDA), rescaled to this model.
Generate estimated from median of 10 measured decode points for this device (llama.cpp CUDA), rescaled to this model.
First token 4K-token prompt ÷ estimated prefill speed. Excludes model load and queueing.

Caveats

Fixed recurrent/convolution state is unknown and not included.
Text-only budget. A downloadable file does not prove your runtime supports this architecture.

AA 7.2 AA estimate LFM2.5-8B-A1B · Arena unknown · Model page

NVIDIA Nemotron 3 Nano 4B

nvidia/NVIDIA-Nemotron-3-Nano-4B-BF16

2026-03-164B dense

AA index4.8rank #75

Arena—

Best file & memory

Q8_03.9 GiB fileFits · conditional

5.1 of 24.0 GiB+8K ctx · max ≈256K

Prompt~20,164 t/sEstimated

Generate~150 t/sEstimated

First token · 4K prompt~203 msEstimated

Memory budget

Weights file	3.94 GiB	unsloth/NVIDIA-Nemotron-3-Nano-4B-GGUF HF files
KV cache · 8K	0.13 GiB	FP16 cache from config.json Estimated
Runtime allowance	1.00 GiB	planning assumption Estimated
Recurrent state	—	Unknown
Accounted	5.07 GiB	of 24.0 GiB usable (100% of 24 GB) · 0.5 GiB spare margin required

Max context in free memory: ≈256K tokens (native 256K).

Other GGUF files

Q4_K_M 2.6 GiB · fits · files
Q5_K_M 2.9 GiB · fits · files

Where the speeds come from

Prompt estimated from median of 20 measured prefill points for this device (llama.cpp CUDA), rescaled to this model.
Generate estimated from median of 24 measured decode points for this device (llama.cpp CUDA), rescaled to this model.
First token 4K-token prompt ÷ estimated prefill speed. Excludes model load and queueing.

Caveats

Fixed recurrent/convolution state is unknown and not included.
Text-only budget. A downloadable file does not prove your runtime supports this architecture.

AA 4.8 NVIDIA Nemotron 3 Nano 4B · Arena unknown · Model page

LFM2.5-24B-A2B

LiquidAI/LFM2.5-24B-A2B

2026-10-1324B · 2B active

AA index—unranked

Arena—

Best file & memory

Q5_K_M15.9 GiB fileFits · conditional

17.0 of 24.0 GiB+8K ctx · max ≈128K

Prompt~16,583 t/sEstimated

Generate~341 t/sEstimated

First token · 4K prompt~247 msEstimated

Memory budget

Weights file	15.86 GiB	LiquidAI/LFM2.5-24B-A2B-GGUF HF files
KV cache · 8K	0.16 GiB	FP16 cache from config.json Estimated
Runtime allowance	1.00 GiB	planning assumption Estimated
Recurrent state	—	Unknown
Accounted	17.02 GiB	of 24.0 GiB usable (100% of 24 GB) · 0.5 GiB spare margin required

Max context in free memory: ≈128K tokens (native 128K).

Other GGUF files

Q4_K_M 13.5 GiB · fits · files
Q8_0 23.7 GiB · too big · files
FP16 44.7 GiB · too big · files

Where the speeds come from

Prompt estimated from median of 4 measured prefill points for this device (llama.cpp CUDA), rescaled to this model.
Generate estimated from median of 10 measured decode points for this device (llama.cpp CUDA), rescaled to this model.
First token 4K-token prompt ÷ estimated prefill speed. Excludes model load and queueing.

Caveats

Fixed recurrent/convolution state is unknown and not included.
Text-only budget. A downloadable file does not prove your runtime supports this architecture.

AA index unknown · Arena unknown · Model page

MiniCPM-V 4.7 35B-A3B

openbmb/MiniCPM-V-4.7-35B-A3B

2026-10-1235B · 3B active

AA index—unranked

Arena—

Best file & memory

Q4_K_M19.7 GiB fileFits · conditional

20.9 of 24.0 GiB+8K ctx · max ≈143K

Prompt~11,055 t/sEstimated

Generate~267 t/sEstimated

First token · 4K prompt~371 msEstimated

Memory budget

Weights file	19.71 GiB	openbmb/MiniCPM-V-4.7-35B-A3B-gguf HF files
KV cache · 8K	0.16 GiB	FP16 cache from config.json Estimated
Runtime allowance	1.00 GiB	planning assumption Estimated
Recurrent state	—	Unknown
Accounted	20.87 GiB	of 24.0 GiB usable (100% of 24 GB) · 0.5 GiB spare margin required

Max context in free memory: ≈143K tokens (native 256K).

Other GGUF files

Q5_K_M 23.0 GiB · too big · files
Q8_0 34.4 GiB · too big · files
FP16 64.6 GiB · too big · files

Where the speeds come from

Prompt estimated from median of 4 measured prefill points for this device (llama.cpp CUDA), rescaled to this model.
Generate estimated from median of 10 measured decode points for this device (llama.cpp CUDA), rescaled to this model.
First token 4K-token prompt ÷ estimated prefill speed. Excludes model load and queueing.

Caveats

Fixed recurrent/convolution state is unknown and not included.
Text-only budget. A downloadable file does not prove your runtime supports this architecture.

AA index unknown · Arena unknown · Model page

Mellum2.1-12B-A2.5B-Thinking

JetBrains/Mellum2.1-12B-A2.5B-Thinking

2026-10-0812B · 2.5B active

AA index—unranked

Arena—

Best file & memory

Q8_012.0 GiB fileFits

13.2 of 24.0 GiB8K ctx · max ≈128K

Prompt~13,266 t/sEstimated

Generate~180 t/sEstimated

First token · 4K prompt~309 msEstimated

Memory budget

Weights file	12.04 GiB	JetBrains/Mellum2.1-12B-A2.5B-Thinking-GGUF HF files
KV cache · 8K	0.15 GiB	FP16 cache from config.json Estimated
Runtime allowance	1.00 GiB	planning assumption Estimated
Accounted	13.19 GiB	of 24.0 GiB usable (100% of 24 GB) · 0.5 GiB spare margin required

Max context in free memory: ≈128K tokens (native 128K).

Other GGUF files

Q4_K_M 7.5 GiB · fits · files
Q5_K_M 9.3 GiB · fits · files

Where the speeds come from

Prompt estimated from median of 4 measured prefill points for this device (llama.cpp CUDA), rescaled to this model.
Generate estimated from median of 10 measured decode points for this device (llama.cpp CUDA), rescaled to this model.
First token 4K-token prompt ÷ estimated prefill speed. Excludes model load and queueing.

Caveats

Text-only budget. A downloadable file does not prove your runtime supports this architecture.

AA index unknown · Arena unknown · Model page

MiMo-V2.6-Distill-Qwen-9B

XiaomiMiMo/MiMo-V2.6-Distill-Qwen-9B

2026-09-229B dense

AA index—unranked

Arena—

Best file & memory

Q8_08.9 GiB fileFits · conditional

10.1 of 24.0 GiB+8K ctx · max ≈256K

Prompt~8,894 t/sEstimated

Generate~66.4 t/sEstimated

First token · 4K prompt~461 msEstimated

Memory budget

Weights file	8.89 GiB	bartowski/MiMo-V2.6-Distill-Qwen-9B-GGUF HF files
KV cache · 8K	0.25 GiB	FP16 cache from config.json Estimated
Runtime allowance	1.00 GiB	planning assumption Estimated
Recurrent state	—	Unknown
Accounted	10.14 GiB	of 24.0 GiB usable (100% of 24 GB) · 0.5 GiB spare margin required

Max context in free memory: ≈256K tokens (native 256K).

Other GGUF files

Q4_K_M 5.4 GiB · fits · files
Q5_K_M 6.4 GiB · fits · files
FP16 16.7 GiB · fits · files

Where the speeds come from

Prompt estimated from median of 20 measured prefill points for this device (llama.cpp CUDA), rescaled to this model.
Generate estimated from median of 24 measured decode points for this device (llama.cpp CUDA), rescaled to this model.
First token 4K-token prompt ÷ estimated prefill speed. Excludes model load and queueing.

Caveats

Fixed recurrent/convolution state is unknown and not included.
Text-only budget. A downloadable file does not prove your runtime supports this architecture.

AA index unknown · Arena unknown · Model page

K2-Horizon-0.9B

IFM/K2-Horizon-0.9B

2026-09-03size unknown

AA index—unranked

Arena—

Best file & memory

Q8_01.1 GiB fileFits

2.5 of 24.0 GiB8K ctx · max ≈128K

Prompt—Unknown

Generate~552 t/sEstimated

First token · 4K prompt—Unknown

Memory budget

Weights file	1.07 GiB	IFM/K2-Horizon-0.9B-GGUF HF files
KV cache · 8K	0.44 GiB	FP16 cache from config.json Estimated
Runtime allowance	1.00 GiB	planning assumption Estimated
Accounted	2.51 GiB	of 24.0 GiB usable (100% of 24 GB) · 0.5 GiB spare margin required

Max context in free memory: ≈128K tokens (native 128K).

Other GGUF files

Q4_K_M 0.6 GiB · fits · files
Q5_K_M 0.7 GiB · fits · files
FP16 2.0 GiB · fits · files

Where the speeds come from

Prompt unknown: no measured report, and no comparable vendor FP16 figure or device calibration.
Generate estimated from median of 24 measured decode points for this device (llama.cpp CUDA), rescaled to this model.
First token unknown: needs a prefill value.

Caveats

Text-only budget. A downloadable file does not prove your runtime supports this architecture.

AA index unknown · Arena unknown · Model page

Ornith-1.5-9B

ornith-ai/Ornith-1.5-9B

2026-08 (day unknown)9B dense

AA index—unranked

Arena—

Best file & memory

Q8_09.1 GiB fileFits · conditional

10.4 of 24.0 GiB+8K ctx · max ≈256K

Prompt~8,894 t/sEstimated

Generate~64.8 t/sEstimated

First token · 4K prompt~461 msEstimated

Memory budget

Weights file	9.11 GiB	ornith-ai/Ornith-1.5-9B-GGUF HF files
KV cache · 8K	0.25 GiB	FP16 cache from config.json Estimated
Runtime allowance	1.00 GiB	planning assumption Estimated
Recurrent state	—	Unknown
Accounted	10.36 GiB	of 24.0 GiB usable (100% of 24 GB) · 0.5 GiB spare margin required

Max context in free memory: ≈256K tokens (native 256K).

Other GGUF files

Q4_K_M 5.4 GiB · fits · files
Q5_K_M 6.2 GiB · fits · files
FP16 16.7 GiB · fits · files

Where the speeds come from

Prompt estimated from median of 20 measured prefill points for this device (llama.cpp CUDA), rescaled to this model.
Generate estimated from median of 24 measured decode points for this device (llama.cpp CUDA), rescaled to this model.
First token 4K-token prompt ÷ estimated prefill speed. Excludes model load and queueing.

Caveats

Fixed recurrent/convolution state is unknown and not included.
Text-only budget. A downloadable file does not prove your runtime supports this architecture.

AA index unknown · Arena unknown · Model page

Qwen3.6-27B

Qwen/Qwen3.6-27B

2026-04-2227B dense

AA index—unranked

Arena—

Best file & memory

Q5_K_M18.2 GiB fileFits · conditional

19.7 of 24.0 GiB+8K ctx · max ≈69K

Prompt~2,965 t/sEstimated

Generate~32.5 t/sEstimated

First token · 4K prompt~1.4 sEstimated

Memory budget

Weights file	18.17 GiB	unsloth/Qwen3.6-27B-GGUF HF files
KV cache · 8K	0.50 GiB	FP16 cache from config.json Estimated
Runtime allowance	1.00 GiB	planning assumption Estimated
Recurrent state	—	Unknown
Accounted	19.67 GiB	of 24.0 GiB usable (100% of 24 GB) · 0.5 GiB spare margin required

Max context in free memory: ≈69K tokens (native 256K).

Other GGUF files

Q4_K_M 15.7 GiB · fits · files
Q8_0 26.6 GiB · too big · files

Where the speeds come from

Prompt estimated from median of 20 measured prefill points for this device (llama.cpp CUDA), rescaled to this model.
Generate estimated from median of 24 measured decode points for this device (llama.cpp CUDA), rescaled to this model.
First token 4K-token prompt ÷ estimated prefill speed. Excludes model load and queueing.

Caveats

Fixed recurrent/convolution state is unknown and not included.
Text-only budget. A downloadable file does not prove your runtime supports this architecture.

AA index unknown · Arena unknown · Model page

Qwen3.5-0.8B

Qwen/Qwen3.5-0.8B

2026-03-020.8B dense

AA index—unranked

Arena—

Best file & memory

Q8_00.8 GiB fileFits · conditional

1.8 of 24.0 GiB+8K ctx · max ≈256K

Prompt~100,063 t/sEstimated

Generate~781 t/sEstimated

First token · 4K prompt~41 msEstimated

Memory budget

Weights file	0.76 GiB	unsloth/Qwen3.5-0.8B-GGUF HF files
KV cache · 8K	0.09 GiB	FP16 cache from config.json Estimated
Runtime allowance	1.00 GiB	planning assumption Estimated
Recurrent state	—	Unknown
Accounted	1.85 GiB	of 24.0 GiB usable (100% of 24 GB) · 0.5 GiB spare margin required

Max context in free memory: ≈256K tokens (native 256K).

Other GGUF files

Q4_K_M 0.5 GiB · fits · files
Q5_K_M 0.5 GiB · fits · files
FP16 1.4 GiB · fits · files

Where the speeds come from

Prompt estimated from median of 20 measured prefill points for this device (llama.cpp CUDA), rescaled to this model.
Generate estimated from median of 24 measured decode points for this device (llama.cpp CUDA), rescaled to this model.
First token 4K-token prompt ÷ estimated prefill speed. Excludes model load and queueing.

Caveats

Fixed recurrent/convolution state is unknown and not included.
Text-only budget. A downloadable file does not prove your runtime supports this architecture.

AA index unknown · Arena unknown · Model page

Qwen3.5-2B

Qwen/Qwen3.5-2B

2026-03-022B dense

AA index—unranked

Arena—

Best file & memory

Q8_01.9 GiB fileFits · conditional

3.0 of 24.0 GiB+8K ctx · max ≈256K

Prompt~40,025 t/sEstimated

Generate~315 t/sEstimated

First token · 4K prompt~102 msEstimated

Memory budget

Weights file	1.87 GiB	unsloth/Qwen3.5-2B-GGUF HF files
KV cache · 8K	0.09 GiB	FP16 cache from config.json Estimated
Runtime allowance	1.00 GiB	planning assumption Estimated
Recurrent state	—	Unknown
Accounted	2.97 GiB	of 24.0 GiB usable (100% of 24 GB) · 0.5 GiB spare margin required

Max context in free memory: ≈256K tokens (native 256K).

Other GGUF files

Q4_K_M 1.2 GiB · fits · files
Q5_K_M 1.3 GiB · fits · files
FP16 3.5 GiB · fits · files

Where the speeds come from

Prompt estimated from median of 20 measured prefill points for this device (llama.cpp CUDA), rescaled to this model.
Generate estimated from median of 24 measured decode points for this device (llama.cpp CUDA), rescaled to this model.
First token 4K-token prompt ÷ estimated prefill speed. Excludes model load and queueing.

Caveats

Fixed recurrent/convolution state is unknown and not included.
Text-only budget. A downloadable file does not prove your runtime supports this architecture.

AA index unknown · Arena unknown · Model page

Qwen3.5-27B

Qwen/Qwen3.5-27B

2026-02-2427B dense

AA index—unranked

Arena—

Best file & memory

Q5_K_M18.3 GiB fileFits · conditional

19.8 of 24.0 GiB+8K ctx · max ≈68K

Prompt~2,965 t/sEstimated

Generate~32.3 t/sEstimated

First token · 4K prompt~1.4 sEstimated

Memory budget

Weights file	18.26 GiB	unsloth/Qwen3.5-27B-GGUF HF files
KV cache · 8K	0.50 GiB	FP16 cache from config.json Estimated
Runtime allowance	1.00 GiB	planning assumption Estimated
Recurrent state	—	Unknown
Accounted	19.76 GiB	of 24.0 GiB usable (100% of 24 GB) · 0.5 GiB spare margin required

Max context in free memory: ≈68K tokens (native 256K).

Other GGUF files

Q4_K_M 15.6 GiB · fits · files
Q8_0 26.6 GiB · too big · files

Where the speeds come from

Prompt estimated from median of 20 measured prefill points for this device (llama.cpp CUDA), rescaled to this model.
Generate estimated from median of 24 measured decode points for this device (llama.cpp CUDA), rescaled to this model.
First token 4K-token prompt ÷ estimated prefill speed. Excludes model load and queueing.

Caveats

Fixed recurrent/convolution state is unknown and not included.
Text-only budget. A downloadable file does not prove your runtime supports this architecture.

AA index unknown · Arena unknown · Model page

Qwen3.5-35B-A3B

Qwen/Qwen3.5-35B-A3B

2026-02-2435B · 3B active

AA index—unranked

Arena—

Best file & memory

Q4_K_M20.5 GiB fileFits · conditional

21.7 of 24.0 GiB+8K ctx · max ≈102K

Prompt~11,055 t/sEstimated

Generate~256 t/sEstimated

First token · 4K prompt~371 msEstimated

Memory budget

Weights file	20.50 GiB	unsloth/Qwen3.5-35B-A3B-GGUF HF files
KV cache · 8K	0.16 GiB	FP16 cache from config.json Estimated
Runtime allowance	1.00 GiB	planning assumption Estimated
Recurrent state	—	Unknown
Accounted	21.66 GiB	of 24.0 GiB usable (100% of 24 GB) · 0.5 GiB spare margin required

Max context in free memory: ≈102K tokens (native 256K).

Other GGUF files

Q5_K_M 24.4 GiB · too big · files
Q8_0 34.4 GiB · too big · files

Where the speeds come from

Prompt estimated from median of 4 measured prefill points for this device (llama.cpp CUDA), rescaled to this model.
Generate estimated from median of 10 measured decode points for this device (llama.cpp CUDA), rescaled to this model.
First token 4K-token prompt ÷ estimated prefill speed. Excludes model load and queueing.

Caveats

Fixed recurrent/convolution state is unknown and not included.
Text-only budget. A downloadable file does not prove your runtime supports this architecture.

AA index unknown · Arena unknown · Model page

May fit: KV-cache size unknown 9

Weights and runtime fit, but the cache for this attention type is not established. Ranked by AA index.

K2-Horizon-MoVA-36B-A4B

IFM/K2-Horizon-MoVA-36B-A4B

2026-09-0336B · 4B active

AA index25.3rank #22

Arena—

Best file & memory

Q4_K_M20.8 GiB fileCache unknown

21.8 of 24.0 GiB+8K ctx

Prompt~8,291 t/sEstimated

Generate~195 t/sEstimated

First token · 4K prompt~494 msEstimated

Memory budget

Weights file	20.83 GiB	IFM/K2-Horizon-MoVA-36B-A4B-GGUF HF files
KV cache · 8K	—	Unknown
Runtime allowance	1.00 GiB	planning assumption Estimated
Recurrent state	—	Unknown
Accounted	21.83 GiB	of 24.0 GiB usable (100% of 24 GB) · 0.5 GiB spare margin required

Max context in free memory: unknown (native 512K).

Other GGUF files

Q5_K_M 24.6 GiB · too big · files
Q8_0 37.1 GiB · too big · files

Where the speeds come from

Prompt estimated from median of 4 measured prefill points for this device (llama.cpp CUDA), rescaled to this model.
Generate estimated from median of 10 measured decode points for this device (llama.cpp CUDA), rescaled to this model.
First token 4K-token prompt ÷ estimated prefill speed. Excludes model load and queueing.

Caveats

KV-cache size could not be established from the config (compressed or backend-specific attention).
Fixed recurrent/convolution state is unknown and not included.
Text-only budget. A downloadable file does not prove your runtime supports this architecture.

AA 25.3 K2 Horizon MoVA 36B A4B · Arena unknown · Model page

Gemma 4 26B A4B (it)

google/gemma-4-26B-A4B-it

2026-04-0225B · 3.8B active

AA index16.7 AA estimaterank #39

Arena1,434

Best file & memory

Q4_K_M15.6 GiB fileCache unknown

16.6 of 24.0 GiB+8K ctx

Prompt~8,728 t/sEstimated

Generate~191 t/sEstimated

First token · 4K prompt~469 msEstimated

Memory budget

Weights file	15.64 GiB	lmstudio-community/gemma-4-26B-A4B-it-GGUF HF files
KV cache · 8K	—	Unknown
Runtime allowance	1.00 GiB	planning assumption Estimated
Recurrent state	—	Unknown
Accounted	16.64 GiB	of 24.0 GiB usable (100% of 24 GB) · 0.5 GiB spare margin required

Max context in free memory: unknown (native 256K).

Other GGUF files

Q8_0 25.0 GiB · too big · files

Where the speeds come from

Prompt estimated from median of 4 measured prefill points for this device (llama.cpp CUDA), rescaled to this model.
Generate estimated from median of 10 measured decode points for this device (llama.cpp CUDA), rescaled to this model.
First token 4K-token prompt ÷ estimated prefill speed. Excludes model load and queueing.

Caveats

KV-cache size could not be established from the config (compressed or backend-specific attention).
Fixed recurrent/convolution state is unknown and not included.
Text-only budget. A downloadable file does not prove your runtime supports this architecture.

AA 16.7 AA estimate Gemma 4 26B A4B (Reasoning) · Arena 1,434 · Model page

Gemma 4 31B (it)

google/gemma-4-31B-it

2026-04-0231B dense

AA index14.7rank #42

Arena1,443

Best file & memory

Q5_K_M20.2 GiB fileCache unknown

21.2 of 24.0 GiB+8K ctx

Prompt~2,607 t/sEstimated

Generate~29.3 t/sEstimated

First token · 4K prompt~1.6 sEstimated

Memory budget

Weights file	20.17 GiB	unsloth/gemma-4-31B-it-GGUF HF files
KV cache · 8K	—	Unknown
Runtime allowance	1.00 GiB	planning assumption Estimated
Recurrent state	—	Unknown
Accounted	21.17 GiB	of 24.0 GiB usable (100% of 24 GB) · 0.5 GiB spare margin required

Max context in free memory: unknown (native 256K).

Other GGUF files

Q4_K_M 17.1 GiB · cache unknown · files
Q8_0 30.4 GiB · too big · files

Where the speeds come from

Prompt estimated from median of 20 measured prefill points for this device (llama.cpp CUDA), rescaled to this model.
Generate estimated from median of 24 measured decode points for this device (llama.cpp CUDA), rescaled to this model.
First token 4K-token prompt ÷ estimated prefill speed. Excludes model load and queueing.

Caveats

KV-cache size could not be established from the config (compressed or backend-specific attention).
Fixed recurrent/convolution state is unknown and not included.
Text-only budget. A downloadable file does not prove your runtime supports this architecture.

AA 14.7 Gemma 4 31B (Reasoning) · Arena 1,443 · Model page

Gemma 4 12B (it)

google/gemma-4-12B-it

2026-06-0312B dense

AA index14.2 AA estimaterank #44

Arena—

Best file & memory

Q8_011.8 GiB fileCache unknown

12.8 of 24.0 GiB+8K ctx

Prompt~6,699 t/sEstimated

Generate~50 t/sEstimated

First token · 4K prompt~611 msEstimated

Memory budget

Weights file	11.80 GiB	bartowski/gemma-4-12B-it-GGUF HF files
KV cache · 8K	—	Unknown
Runtime allowance	1.00 GiB	planning assumption Estimated
Recurrent state	—	Unknown
Accounted	12.80 GiB	of 24.0 GiB usable (100% of 24 GB) · 0.5 GiB spare margin required

Max context in free memory: unknown (native 256K).

Other GGUF files

Q4_K_M 7.1 GiB · cache unknown · files
Q5_K_M 8.2 GiB · cache unknown · files

Where the speeds come from

Prompt estimated from median of 20 measured prefill points for this device (llama.cpp CUDA), rescaled to this model.
Generate estimated from median of 24 measured decode points for this device (llama.cpp CUDA), rescaled to this model.
First token 4K-token prompt ÷ estimated prefill speed. Excludes model load and queueing.

Caveats

KV-cache size could not be established from the config (compressed or backend-specific attention).
Fixed recurrent/convolution state is unknown and not included.
Text-only budget. A downloadable file does not prove your runtime supports this architecture.

AA 14.2 AA estimate Gemma 4 12B (Reasoning) · Arena unknown · Model page

Ling-3.0-tiny

inclusionAI/Ling-3.0-tiny

2026-08 (day unknown)7.9B · 1.3B active

AA index11.1rank #57

Arena—

Best file & memory

Q8_07.8 GiB fileCache unknown

8.8 of 24.0 GiB+8K ctx

Prompt~25,512 t/sEstimated

Generate~350 t/sEstimated

First token · 4K prompt~161 msEstimated

Memory budget

Weights file	7.83 GiB	inclusionAI/Ling-3.0-tiny-GGUF HF files
KV cache · 8K	—	Unknown
Runtime allowance	1.00 GiB	planning assumption Estimated
Recurrent state	—	Unknown
Accounted	8.83 GiB	of 24.0 GiB usable (100% of 24 GB) · 0.5 GiB spare margin required

Max context in free memory: unknown (native 128K).

Other GGUF files

Q4_K_M 4.5 GiB · cache unknown · files
Q5_K_M 5.2 GiB · cache unknown · files
FP16 14.7 GiB · cache unknown · files

Where the speeds come from

Prompt estimated from median of 4 measured prefill points for this device (llama.cpp CUDA), rescaled to this model.
Generate estimated from median of 10 measured decode points for this device (llama.cpp CUDA), rescaled to this model.
First token 4K-token prompt ÷ estimated prefill speed. Excludes model load and queueing.

Caveats

KV-cache size could not be established from the config (compressed or backend-specific attention).
Fixed recurrent/convolution state is unknown and not included.
Text-only budget. A downloadable file does not prove your runtime supports this architecture.

AA 11.1 Ling 3.0 Tiny · Arena unknown · Model page

DiffusionGemma 26B A4B

google/diffusiongemma-26B-A4B-it

2026-06-1025B · 3.8B active

AA index9.5 AA estimaterank #62

Arena—

Best file & memory

Q5_K_M17.8 GiB fileCache unknown

18.8 of 24.0 GiB+8K ctx

Prompt~8,728 t/sEstimated

Generate~168 t/sEstimated

First token · 4K prompt~469 msEstimated

Memory budget

Weights file	17.83 GiB	unsloth/diffusiongemma-26B-A4B-it-GGUF HF files
KV cache · 8K	—	Unknown
Runtime allowance	1.00 GiB	planning assumption Estimated
Recurrent state	—	Unknown
Accounted	18.83 GiB	of 24.0 GiB usable (100% of 24 GB) · 0.5 GiB spare margin required

Max context in free memory: unknown (native 256K).

Other GGUF files

Q4_K_M 15.7 GiB · cache unknown · files
Q8_0 25.0 GiB · too big · files

Where the speeds come from

Prompt estimated from median of 4 measured prefill points for this device (llama.cpp CUDA), rescaled to this model.
Generate estimated from median of 10 measured decode points for this device (llama.cpp CUDA), rescaled to this model.
First token 4K-token prompt ÷ estimated prefill speed. Excludes model load and queueing.

Caveats

KV-cache size could not be established from the config (compressed or backend-specific attention).
Fixed recurrent/convolution state is unknown and not included.
Text-only budget. A downloadable file does not prove your runtime supports this architecture.

AA 9.5 AA estimate DiffusionGemma 26B A4B · Arena unknown · Model page

Gemma 4 E4B (it)

google/gemma-4-E4B-it

2026-04-028B dense

AA index8.9rank #66

Arena—

Best file & memory

Q8_07.6 GiB fileCache unknown

8.6 of 24.0 GiB+8K ctx

Prompt~10,006 t/sEstimated

Generate~77.3 t/sEstimated

First token · 4K prompt~409 msEstimated

Memory budget

Weights file	7.63 GiB	unsloth/gemma-4-E4B-it-GGUF HF files
KV cache · 8K	—	Unknown
Runtime allowance	1.00 GiB	planning assumption Estimated
Recurrent state	—	Unknown
Accounted	8.63 GiB	of 24.0 GiB usable (100% of 24 GB) · 0.5 GiB spare margin required

Max context in free memory: unknown (native 128K).

Other GGUF files

Q4_K_M 4.6 GiB · cache unknown · files
Q5_K_M 5.1 GiB · cache unknown · files

Where the speeds come from

Prompt estimated from median of 20 measured prefill points for this device (llama.cpp CUDA), rescaled to this model.
Generate estimated from median of 24 measured decode points for this device (llama.cpp CUDA), rescaled to this model.
First token 4K-token prompt ÷ estimated prefill speed. Excludes model load and queueing.

Caveats

KV-cache size could not be established from the config (compressed or backend-specific attention).
Fixed recurrent/convolution state is unknown and not included.
Text-only budget. A downloadable file does not prove your runtime supports this architecture.

AA 8.9 Gemma 4 E4B (Reasoning) · Arena unknown · Model page

Gemma 4 E2B

google/gemma-4-E2B-it

2026-04-025.1B dense

AA index7.8 AA estimaterank #72

Arena—

Best file & memory

Q8_04.7 GiB fileCache unknown

5.7 of 24.0 GiB+8K ctx

Prompt~15,696 t/sEstimated

Generate~126 t/sEstimated

First token · 4K prompt~261 msEstimated

Memory budget

Weights file	4.70 GiB	unsloth/gemma-4-E2B-it-GGUF HF files
KV cache · 8K	—	Unknown
Runtime allowance	1.00 GiB	planning assumption Estimated
Recurrent state	—	Unknown
Accounted	5.70 GiB	of 24.0 GiB usable (100% of 24 GB) · 0.5 GiB spare margin required

Max context in free memory: unknown (native 128K).

Other GGUF files

Q4_K_M 2.9 GiB · cache unknown · files
Q5_K_M 3.1 GiB · cache unknown · files

Where the speeds come from

Prompt estimated from median of 20 measured prefill points for this device (llama.cpp CUDA), rescaled to this model.
Generate estimated from median of 24 measured decode points for this device (llama.cpp CUDA), rescaled to this model.
First token 4K-token prompt ÷ estimated prefill speed. Excludes model load and queueing.

Caveats

KV-cache size could not be established from the config (compressed or backend-specific attention).
Fixed recurrent/convolution state is unknown and not included.
Text-only budget. A downloadable file does not prove your runtime supports this architecture.

AA 7.8 AA estimate Gemma 4 E2B (Reasoning) · Arena unknown · Model page

GLM-4.7-Flash

zai-org/GLM-4.7-Flash

2026-01 (day unknown)30B · 3B active

AA index—unranked

Arena—

Best file & memory

Q5_K_M19.9 GiB fileCache unknown

20.9 of 24.0 GiB+8K ctx

Prompt~11,055 t/sEstimated

Generate~226 t/sEstimated

First token · 4K prompt~371 msEstimated

Memory budget

Weights file	19.94 GiB	unsloth/GLM-4.7-Flash-GGUF HF files
KV cache · 8K	—	Unknown
Runtime allowance	1.00 GiB	planning assumption Estimated
Recurrent state	—	Unknown
Accounted	20.94 GiB	of 24.0 GiB usable (100% of 24 GB) · 0.5 GiB spare margin required

Max context in free memory: unknown (native 198K).

Other GGUF files

Q4_K_M 17.1 GiB · cache unknown · files
Q8_0 29.7 GiB · too big · files

Where the speeds come from

Prompt estimated from median of 4 measured prefill points for this device (llama.cpp CUDA), rescaled to this model.
Generate estimated from median of 10 measured decode points for this device (llama.cpp CUDA), rescaled to this model.
First token 4K-token prompt ÷ estimated prefill speed. Excludes model load and queueing.

Caveats

KV-cache size could not be established from the config (compressed or backend-specific attention).
Fixed recurrent/convolution state is unknown and not included.
Text-only budget. A downloadable file does not prove your runtime supports this architecture.

AA index unknown · Arena unknown · Model page

42 models too big for RTX 4090

GLM-5.3Q8_0 needs 748 GiB+ (KV unknown)
GLM-5.3-FlashQ8_0 needs 319 GiB+ (KV unknown)
Ling 3.1 FlashQ4_K_M needs 335 GiB+ (KV unknown)
Qwen3.8-2.4T-A95BQ8_0 needs 2,424 GiB
Qwen3.8-Flash-NextQ8_0 needs 177 GiB+ (KV unknown)
Muse Spark Open (816B-A42B)Q8_0 needs 813 GiB+ (KV unknown)
Motif 3Q4_K_M needs 185 GiB+ (KV unknown)
K2-Horizon-375B-A23BQ4_K_M needs 225 GiB
Reflection Beam 1.0 501B-A23BQ8_0 needs 498 GiB
MiniMax-M3Q4_K_M needs 245 GiB+ (KV unknown)
Nex-N2-Pro (Qwen3.5-397B-A17B based)Q4_K_M needs 227 GiB
NVIDIA Nemotron 3.5 Super VL 120B-A12BQ8_0 needs 124 GiB
Kimi K2.7 CodeQ4_K_M needs 579 GiB+ (KV unknown)
Inkling-SmallQ8_0 needs 263 GiB
Hy3Q4_K_M needs 174 GiB
InklingQ8_0 needs 800 GiB
Solar Open 2 (250B-A15B)Q8_0 needs 250 GiB
Ling-3.0-flash-VLQ4_K_M needs 74.8 GiB+ (KV unknown)
NVIDIA Nemotron 3 Ultra 550B-A55BQ4_K_M needs 323 GiB
Ling-3.0-flash-FinQ4_K_M needs 74.0 GiB+ (KV unknown)
Qwen3.5-397B-A17BQ4_K_M needs 229 GiB
NVIDIA Nemotron 3.5 Bolt 74B-A9BQ8_0 needs 75.0 GiB
Qwen3.5-122B-A10BQ4_K_M needs 73.0 GiB
Mistral Medium 3.5 (128B dense)Q4_K_M needs 74.0 GiB
Command A+ (05-2026)Q4_K_M needs 125 GiB
NVIDIA Nemotron 3.5 Lightning 30B-A3BQ4_K_M needs 25.3 GiB
HyperNova 60B 2605 (compressed gpt-oss-120b)Q4_K_M needs 39.6 GiB
Nemotron Cascade 2 30B A3BQ4_K_M needs 24.4 GiB
LongCat-Flash-LiteQ4_K_M needs 39.0 GiB+ (KV unknown)
Mistral Small 4 (119B-2603)Q4_K_M needs 69.1 GiB+ (KV unknown)
Trinity-Large-ThinkingQ4_K_M needs 228 GiB
Nemotron 3 Nano Omni 30B A3B ReasoningQ4_K_M needs 24.3 GiB+ (KV unknown)
Qwen3-Coder-NextQ4_K_M needs 46.9 GiB
Sarvam 105BQ4_K_M needs 61.3 GiB+ (KV unknown)
Aleph Alpha Kolibri-1Q4_K_M needs 45.9 GiB
Hy4-previewQ4_K_M needs 437 GiB+ (KV unknown)
Ling-3.0-flashQ4_K_M needs 74.0 GiB+ (KV unknown)
Poolside Laguna S 2.1Q4_K_M needs 91.4 GiB
GLM-5.2Q8_0 needs 748 GiB+ (KV unknown)
Step 3.7 FlashQ8_0 needs 197 GiB+ (KV unknown)
NVIDIA Nemotron 3 Super 120B-A12BQ8_0 needs 121 GiB
MiniMax-M2.7Q4_K_M needs 133 GiB

Specs & sources

Memory	24 GB GDDR6X	source
Bandwidth	1,008 GB/s	source
FP16	165.2 TFLOPS	source
Power	450 W	source

FP16: Peak FP16 Tensor TFLOPS with FP32 Accumulate (dense). Power: Total Graphics Power.

Measured speeds

91 reported observations

On this device. Reported by the linked sources; not reproduced by LocalFit.

Open in speed database Submit yours

Doesn't fit? Rent a GPU

RunpodPlaceholderCloud GPU pods you can rent for running larger models.
Vast.aiPlaceholderGPU rental marketplace with many hosts and GPU types.
LambdaPlaceholderCloud GPU instances for AI training and inference.
TensorDockPlaceholderGPU cloud marketplace with consumer and datacenter GPUs.

Sponsored placeholder links: LocalFit may earn a commission if these become affiliate links. They never influence fit results, estimates or picks.