NVIDIA GeForce RTX 3080 vs NVIDIA GeForce RTX 4060
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)
NVIDIA GeForce RTX 3080
- 10GB
- ampere
- nvidia
- $699
- 135 of 145
- 26 models
NVIDIA GeForce RTX 4060
- 8GB
- ada lovelace
- nvidia
- $299
- 132 of 145
- 0 models
Where NVIDIA GeForce RTX 3080 pulls ahead
- Gemma 2 9B InstructSvsA
- Phi-4BvsC
- CodeGemma 7BSvsA
- Gemma 3 12BAvsB
- Qwen 2.5 14BBvsC
Where NVIDIA GeForce RTX 4060 pulls ahead
- No standout wins.
Language Models74 models
3080
tok/s
Model
tok/s
4060
D
—
Qwen 2.5 32B
32B · Alibaba
—
F
D
—
Skyfall 31B v4.2
31B · TheDrummer
—
F
D
—
Qwen3 30B-A3B
30.5B · Alibaba
—
F
C
—
DeepSeek MoE 16B
16.4B · DeepSeek
—
D
B
—
Rocinante XL 16B v1
16B · TheDrummer
—
C
B
—
Phi-4
14B · Microsoft
—
C
B
—
Qwen 2.5 14B
14B · Alibaba
—
C
A
—
Gemma 3 12B
12B · Google
—
B
A
—
Mistral Nemo 12B
12B · Mistral AI
—
B
A
—
Magnum v4 12B
12B · Anthracite
—
B
A
—
Rocinante 12B v1.1
12B · TheDrummer
—
B
A
—
Mistral Nemo Base 12B
12B · Mistral AI
—
B
S
—
Solar 10.7B
10.7B · Upstage
—
A
S
46
Falcon 3 10B
10B · TII
40
A
S
46
Gemma 2 9B Instruct
9.2B · Google
40
A
S
46
Yi 1.5 9B Chat
9B · 01.AI
40
A
A
46
Gemma 3 MoE 9B
9B · Google
38
B
F
—
Qwen3 235B-A22B
235B · Alibaba
—
F
F
—
Mixtral 8x22B Instruct
141B · Mistral AI
—
F
F
—
Magnum v4 72B
72B · Anthracite
—
F
F
—
Llama 3.1 70B Instruct
70B · Meta
—
F
F
—
Euryale L3.3 70B v2.3
70B · Sao10K
—
F
F
—
Llama 3.1 70B (lorablated)
70B · mlabonne
—
F
F
—
Mixtral 8x7B Instruct
46.7B · Mistral AI
—
F
F
—
Phi-3.5 MoE
41.9B · Microsoft
—
F
D
—
Gemma 3 27B
27B · Google
—
D
D
—
Dolphin Mistral 24B (Venice Edition)
24B · Cognitive Computations
—
D
D
—
Dolphin 3.0 R1 Mistral 24B
24B · Cognitive Computations
—
D
D
—
Cydonia 24B v4.3
24B · TheDrummer
—
D
D
—
Mistral Small 22B
22B · Mistral AI
—
D
D
—
Magnum v4 22B
22B · Anthracite
—
D
S
46
DeepSeek R1 Distill 8B
8B · DeepSeek
40
S
S
46
Llama 3.1 8B Instruct
8B · Meta
40
S
S
46
Granite 3.3 8B
8B · IBM
40
S
S
46
Dolphin 3.0 Llama 3.1 8B
8B · Cognitive Computations
40
S
S
46
NeuralDaredevil 8B (abliterated)
8B · mlabonne
40
S
S
46
Llama 3.1 8B Instruct (abliterated)
8B · mlabonne
40
S
S
46
Stheno L3 8B v3.2
8B · Sao10K
40
S
S
46
Qwen3 8B Base
8B · Alibaba
40
S
S
46
EXAONE 3.5 7.8B
7.8B · LG AI
40
S
S
46
InternLM 2.5 7B
7.7B · Shanghai AI Lab
40
S
S
46
Qwen 2.5 7B Instruct
7.6B · Alibaba
40
S
S
46
Mistral 7B Instruct v0.3
7.3B · Mistral AI
40
S
S
46
Falcon 3 7B
7B · TII
40
S
S
46
OLMo 2 7B
7B · Allen AI
40
S
S
46
OpenChat 3.5 7B
7B · OpenChat
40
S
S
46
OLMoE 1B-7B
6.9B · AI2
40
S
S
46
Yi 1.5 6B Chat
6B · 01.AI
40
S
S
74
Gemma 3 4B
4B · Google
70
S
S
74
Nemotron Mini 4B
4B · NVIDIA
70
S
S
74
Danube 3 4B
4B · H2O.ai
70
S
S
74
Phi-3.5 Mini 3.8B
3.8B · Microsoft
70
S
S
74
Phi-4 Mini 3.8B
3.8B · Microsoft
70
S
S
74
Granite 3.0 3B-A800M
3.4B · IBM
70
S
S
74
Llama 3.2 3B Instruct
3.2B · Meta
70
S
S
74
Qwen 2.5 3B
3B · Alibaba
70
S
S
74
Falcon 3 3B
3B · TII
70
S
S
74
StableLM Zephyr 3B
3B · Stability AI
70
S
S
74
Rocket 3B
3B · Pansophic
70
S
S
74
Gemma 2 2B
2.6B · Google
70
S
S
74
EXAONE 3.5 2.4B
2.4B · LG AI
70
S
S
108
Granite 3.3 2B
2B · IBM
102
S
S
108
SmolLM2 1.7B
1.7B · HuggingFace
102
S
S
108
Qwen 2.5 1.5B
1.5B · Alibaba
102
S
S
108
DeepSeek R1 Distill 1.5B
1.5B · DeepSeek
102
S
S
108
Granite 3.0 1B-A400M
1.3B · IBM
102
S
S
108
Llama 3.2 1B Instruct
1.24B · Meta
102
S
S
108
TinyLlama 1.1B
1.1B · TinyLlama
102
S
S
108
Gemma 3 1B
1B · Google
102
S
S
108
Falcon 3 1B
1B · TII
102
S
S
108
Qwen 2.5 0.5B
0.5B · Alibaba
102
S
S
108
Danube 3 500M
0.5B · H2O.ai
102
S
S
108
SmolLM2 360M
0.36B · HuggingFace
102
S
S
108
SmolLM2 135M
0.135B · HuggingFace
102
S
Code Models17 models
3080
tok/s
Model
tok/s
4060
B
—
Qwen 2.5 Coder 14B
14B · Alibaba
—
C
A
—
Code Llama 13B Instruct
13B · Meta
—
B
S
46
Yi Coder 9B
9B · 01.AI
40
A
S
46
CodeGemma 7B
8.5B · Google
40
A
D
—
Codestral 22B (abliterated)
22B · failspy
—
D
S
46
Qwen 2.5 Coder 7B
7.6B · Alibaba
40
S
S
46
StarCoder2 7B
7B · BigCode
40
S
S
46
Code Llama 7B
7B · Meta
40
S
S
46
DeepSeek Coder 6.7B
6.7B · DeepSeek
40
S
S
74
Qwen 2.5 Coder 3B
3B · Alibaba
70
S
S
74
StarCoder2 3B
3B · BigCode
70
S
S
74
Stable Code 3B
3B · Stability AI
70
S
S
108
CodeGemma 2B
2B · Google
102
S
S
108
Qwen 2.5 Coder 1.5B
1.5B · Alibaba
102
S
S
108
Yi Coder 1.5B
1.5B · 01.AI
102
S
S
108
DeepSeek Coder 1.3B
1.3B · DeepSeek
102
S
S
108
Qwen 2.5 Coder 0.5B
0.5B · Alibaba
102
S
Multimodal & Vision6 models
3080
tok/s
Model
tok/s
4060
Image Generation9 models
3080
tok/s
Model
tok/s
4060
B
68
Stable Diffusion 3 Medium (GGUF)
2.5B · Stability AI
51
C
D
—
FLUX.1 Schnell (GGUF)
12B · Black Forest Labs
—
D
D
—
FLUX.1 Dev (GGUF)
12B · Black Forest Labs
—
D
S
74
Stable Diffusion XL (CoreML)
3.5B · Stability AI
70
S
S
74
SDXL Turbo (GGUF)
3.5B · Stability AI
70
S
S
108
Stable Diffusion 2.1 Base (CoreML)
0.86B · Stability AI / Apple
102
S
S
108
Stable Diffusion 1.5 (CoreML)
0.86B · Runway
102
S
S
108
Stable Diffusion 1.5 (GGUF)
0.86B · Runway / GPUStack
102
S
S
108
Stable Diffusion 2.1 (GGUF)
0.86B · Stability AI
102
S
Speech9 models
3080
tok/s
Model
tok/s
4060
S
108
Whisper Large v3
1.55B · OpenAI
102
S
S
108
Whisper Large v3 Turbo
0.81B · OpenAI
102
S
S
108
Whisper Medium
0.77B · OpenAI
102
S
S
108
Distil-Whisper Large v3
0.76B · HuggingFace
102
S
S
108
Whisper Small
0.24B · OpenAI
102
S
S
108
Whisper Base
0.074B · OpenAI
102
S
S
108
Whisper Base English
0.074B · OpenAI
102
S
S
108
Whisper Tiny English (Quantized)
0.039B · OpenAI
102
S
S
108
Whisper Tiny
0.039B · OpenAI
102
S
Text-to-Speech14 models
3080
tok/s
Model
tok/s
4060
S
108
Kokoro 82M TTS
0.082B · Kokoro
102
S
S
108
Piper TTS - Amy (English)
0.02B · Rhasspy
102
S
S
108
Piper TTS - Lessac (English)
0.02B · Rhasspy
102
S
S
108
Piper TTS - LibriTTS-R (English)
0.02B · Rhasspy
102
S
S
108
Piper TTS - Spanish (MLS)
0.02B · Rhasspy
102
S
S
108
Piper TTS - French (Siwis)
0.02B · Rhasspy
102
S
S
108
Piper TTS - German (Thorsten)
0.02B · Rhasspy
102
S
S
108
Piper TTS - Chinese (Huayan)
0.02B · Rhasspy
102
S
S
108
Piper TTS - Japanese (Kokoro)
0.02B · Rhasspy
102
S
S
108
Piper TTS - Korean
0.02B · Rhasspy
102
S
S
108
Piper TTS - Russian (Irina)
0.02B · Rhasspy
102
S
S
108
Piper TTS - Portuguese (Faber)
0.02B · Rhasspy
102
S
S
108
Piper TTS - Italian (Riccardo)
0.02B · Rhasspy
102
S
S
108
Piper TTS - Arabic (Kareem)
0.02B · Rhasspy
102
S
Embeddings5 models
3080
tok/s
Model
tok/s
4060
Rerankers2 models
3080
tok/s
Model
tok/s
4060