▣~/runthismodel▍
daemon okbuild 5a3c91d00:00:00Z

Apple M4 Pro vs Apple M4 Ultra

Head-to-head AI inference comparison across 145 popular models. Each model is graded against both cards using its highest-quality quantization that still fits in VRAM. Bigger grade and faster tokens-per-second wins.

Spec
  • VRAM
  • Architecture
  • Vendor
  • MSRP
  • Models running
  • Wins (grade)
Apple M4 Pro
  • 48GB
  • m4
  • apple
  • —
  • 143 of 145
  • 0 models
Apple M4 Ultra
  • 192GB
  • m4
  • apple
  • —
  • 145 of 145
  • 10 models

Where Apple M4 Pro pulls ahead

  • No standout wins.

Where Apple M4 Ultra pulls ahead

Language Models74 models

F
—
Mixtral 8x22B Instruct
141B · Mistral AI
22
S
F
—
Qwen3 235B-A22B
235B · Alibaba
22
B
C
—
Magnum v4 72B
72B · Anthracite
22
S
C
—
Llama 3.1 70B Instruct
70B · Meta
22
S
C
—
Euryale L3.3 70B v2.3
70B · Sao10K
22
S
C
—
Llama 3.1 70B (lorablated)
70B · mlabonne
22
S
A
17
Mixtral 8x7B Instruct
46.7B · Mistral AI
30
S
A
17
Phi-3.5 MoE
41.9B · Microsoft
30
S
S
17
Qwen 2.5 32B
32B · Alibaba
30
S
S
17
Skyfall 31B v4.2
31B · TheDrummer
30
S
S
17
Qwen3 30B-A3B
30.5B · Alibaba
30
S
S
17
Gemma 3 27B
27B · Google
30
S
S
17
Dolphin Mistral 24B (Venice Edition)
24B · Cognitive Computations
30
S
S
17
Dolphin 3.0 R1 Mistral 24B
24B · Cognitive Computations
30
S
S
17
Cydonia 24B v4.3
24B · TheDrummer
30
S
S
17
Mistral Small 22B
22B · Mistral AI
30
S
S
17
Magnum v4 22B
22B · Anthracite
30
S
S
26
DeepSeek MoE 16B
16.4B · DeepSeek
42
S
S
26
Rocinante XL 16B v1
16B · TheDrummer
42
S
S
26
Phi-4
14B · Microsoft
42
S
S
26
Qwen 2.5 14B
14B · Alibaba
42
S
S
26
Gemma 3 12B
12B · Google
42
S
S
26
Mistral Nemo 12B
12B · Mistral AI
42
S
S
26
Magnum v4 12B
12B · Anthracite
42
S
S
26
Rocinante 12B v1.1
12B · TheDrummer
42
S
S
26
Mistral Nemo Base 12B
12B · Mistral AI
42
S
S
26
Solar 10.7B
10.7B · Upstage
42
S
S
38
Falcon 3 10B
10B · TII
54
S
S
38
Gemma 2 9B Instruct
9.2B · Google
54
S
S
38
Yi 1.5 9B Chat
9B · 01.AI
54
S
S
38
Gemma 3 MoE 9B
9B · Google
54
S
S
38
DeepSeek R1 Distill 8B
8B · DeepSeek
54
S
S
38
Llama 3.1 8B Instruct
8B · Meta
54
S
S
38
Granite 3.3 8B
8B · IBM
54
S
S
38
Dolphin 3.0 Llama 3.1 8B
8B · Cognitive Computations
54
S
S
38
NeuralDaredevil 8B (abliterated)
8B · mlabonne
54
S
S
38
Llama 3.1 8B Instruct (abliterated)
8B · mlabonne
54
S
S
38
Stheno L3 8B v3.2
8B · Sao10K
54
S
S
38
Qwen3 8B Base
8B · Alibaba
54
S
S
38
EXAONE 3.5 7.8B
7.8B · LG AI
54
S
S
38
InternLM 2.5 7B
7.7B · Shanghai AI Lab
54
S
S
38
Qwen 2.5 7B Instruct
7.6B · Alibaba
54
S
S
38
Mistral 7B Instruct v0.3
7.3B · Mistral AI
54
S
S
38
Falcon 3 7B
7B · TII
54
S
S
38
OLMo 2 7B
7B · Allen AI
54
S
S
38
OpenChat 3.5 7B
7B · OpenChat
54
S
S
38
OLMoE 1B-7B
6.9B · AI2
54
S
S
38
Yi 1.5 6B Chat
6B · 01.AI
54
S
S
62
Gemma 3 4B
4B · Google
84
S
S
62
Nemotron Mini 4B
4B · NVIDIA
84
S
S
62
Danube 3 4B
4B · H2O.ai
84
S
S
62
Phi-3.5 Mini 3.8B
3.8B · Microsoft
84
S
S
62
Phi-4 Mini 3.8B
3.8B · Microsoft
84
S
S
62
Granite 3.0 3B-A800M
3.4B · IBM
84
S
S
62
Llama 3.2 3B Instruct
3.2B · Meta
84
S
S
62
Qwen 2.5 3B
3B · Alibaba
84
S
S
62
Falcon 3 3B
3B · TII
84
S
S
62
StableLM Zephyr 3B
3B · Stability AI
84
S
S
62
Rocket 3B
3B · Pansophic
84
S
S
62
Gemma 2 2B
2.6B · Google
84
S
S
62
EXAONE 3.5 2.4B
2.4B · LG AI
84
S
S
90
Granite 3.3 2B
2B · IBM
114
S
S
90
SmolLM2 1.7B
1.7B · HuggingFace
114
S
S
90
Qwen 2.5 1.5B
1.5B · Alibaba
114
S
S
90
DeepSeek R1 Distill 1.5B
1.5B · DeepSeek
114
S
S
90
Granite 3.0 1B-A400M
1.3B · IBM
114
S
S
90
Llama 3.2 1B Instruct
1.24B · Meta
114
S
S
90
TinyLlama 1.1B
1.1B · TinyLlama
114
S
S
90
Gemma 3 1B
1B · Google
114
S
S
90
Falcon 3 1B
1B · TII
114
S
S
90
Qwen 2.5 0.5B
0.5B · Alibaba
114
S
S
90
Danube 3 500M
0.5B · H2O.ai
114
S
S
90
SmolLM2 360M
0.36B · HuggingFace
114
S
S
90
SmolLM2 135M
0.135B · HuggingFace
114
S

Code Models17 models

S
17
Codestral 22B (abliterated)
22B · failspy
30
S
S
26
Qwen 2.5 Coder 14B
14B · Alibaba
42
S
S
26
Code Llama 13B Instruct
13B · Meta
42
S
S
38
Yi Coder 9B
9B · 01.AI
54
S
S
38
CodeGemma 7B
8.5B · Google
54
S
S
38
Qwen 2.5 Coder 7B
7.6B · Alibaba
54
S
S
38
StarCoder2 7B
7B · BigCode
54
S
S
38
Code Llama 7B
7B · Meta
54
S
S
38
DeepSeek Coder 6.7B
6.7B · DeepSeek
54
S
S
62
Qwen 2.5 Coder 3B
3B · Alibaba
84
S
S
62
StarCoder2 3B
3B · BigCode
84
S
S
62
Stable Code 3B
3B · Stability AI
84
S
S
90
CodeGemma 2B
2B · Google
114
S
S
90
Qwen 2.5 Coder 1.5B
1.5B · Alibaba
114
S
S
90
Yi Coder 1.5B
1.5B · 01.AI
114
S
S
90
DeepSeek Coder 1.3B
1.3B · DeepSeek
114
S
S
90
Qwen 2.5 Coder 0.5B
0.5B · Alibaba
114
S

Multimodal & Vision6 models

S
38
LLaVA 1.6 7B
7B · LLaVA
54
S
S
62
Phi-3.5 Vision
4.2B · Microsoft
84
S
S
62
PaliGemma 3B
3B · Google
84
S
S
62
Qwen2-VL 2B
2.2B · Alibaba
84
S
S
90
MiniCPM-V 2.6
2B · OpenBMB
114
S
S
90
Moondream 2
1.8B · Moondream
114
S

Image Generation9 models

S
26
FLUX.1 Schnell (GGUF)
12B · Black Forest Labs
42
S
S
26
FLUX.1 Dev (GGUF)
12B · Black Forest Labs
42
S
S
62
Stable Diffusion XL (CoreML)
3.5B · Stability AI
84
S
S
62
SDXL Turbo (GGUF)
3.5B · Stability AI
84
S
S
62
Stable Diffusion 3 Medium (GGUF)
2.5B · Stability AI
84
S
S
90
Stable Diffusion 2.1 Base (CoreML)
0.86B · Stability AI / Apple
114
S
S
90
Stable Diffusion 1.5 (CoreML)
0.86B · Runway
114
S
S
90
Stable Diffusion 1.5 (GGUF)
0.86B · Runway / GPUStack
114
S
S
90
Stable Diffusion 2.1 (GGUF)
0.86B · Stability AI
114
S

Speech9 models

S
90
Whisper Large v3
1.55B · OpenAI
114
S
S
90
Whisper Large v3 Turbo
0.81B · OpenAI
114
S
S
90
Whisper Medium
0.77B · OpenAI
114
S
S
90
Distil-Whisper Large v3
0.76B · HuggingFace
114
S
S
90
Whisper Small
0.24B · OpenAI
114
S
S
90
Whisper Base
0.074B · OpenAI
114
S
S
90
Whisper Base English
0.074B · OpenAI
114
S
S
90
Whisper Tiny English (Quantized)
0.039B · OpenAI
114
S
S
90
Whisper Tiny
0.039B · OpenAI
114
S

Text-to-Speech14 models

S
90
Kokoro 82M TTS
0.082B · Kokoro
114
S
S
90
Piper TTS - Amy (English)
0.02B · Rhasspy
114
S
S
90
Piper TTS - Lessac (English)
0.02B · Rhasspy
114
S
S
90
Piper TTS - LibriTTS-R (English)
0.02B · Rhasspy
114
S
S
90
Piper TTS - Spanish (MLS)
0.02B · Rhasspy
114
S
S
90
Piper TTS - French (Siwis)
0.02B · Rhasspy
114
S
S
90
Piper TTS - German (Thorsten)
0.02B · Rhasspy
114
S
S
90
Piper TTS - Chinese (Huayan)
0.02B · Rhasspy
114
S
S
90
Piper TTS - Japanese (Kokoro)
0.02B · Rhasspy
114
S
S
90
Piper TTS - Korean
0.02B · Rhasspy
114
S
S
90
Piper TTS - Russian (Irina)
0.02B · Rhasspy
114
S
S
90
Piper TTS - Portuguese (Faber)
0.02B · Rhasspy
114
S
S
90
Piper TTS - Italian (Riccardo)
0.02B · Rhasspy
114
S
S
90
Piper TTS - Arabic (Kareem)
0.02B · Rhasspy
114
S

Embeddings5 models

S
90
BGE Large EN v1.5
0.335B · BAAI
114
S
S
90
Nomic Embed Text v1.5
0.137B · Nomic AI
114
S
S
90
BGE Small EN v1.5
0.033B · BAAI
114
S
S
90
Snowflake Arctic Embed S
0.033B · Snowflake
114
S
S
90
all-MiniLM-L6-v2
0.023B · Sentence Transformers
114
S

Rerankers2 models

S
90
BGE Reranker v2 M3
0.568B · BAAI
114
S
S
90
Jina Reranker Tiny EN
0.033B · Jina AI
114
S
© runthismodel · 2026privacytermsdisclaimerchangelogembed badgerunpodvast.aihuggingfaceollamalm-studiomade for the people who actually read GGUF metadata
 ┌─┐                ╔══╗     ╔══╗
 │░│  RUN  THIS  M  ║▓▓║ DEL ║▓▓║
 └─┘                ╚══╝     ╚══╝