Model Comparisons

Side-by-side inference performance comparisons across AI models. Compare throughput, cost per token, memory requirements, and quality benchmarks.

Reasoning

DeepSeek R1vsQwen 3 235B

DeepSeek R1vsDeepSeek V3

DeepSeek R1vsDeepSeek R1 Distill 70B

DeepSeek R1 Distill 70BvsQwen 3 32B

DeepSeek R1 Distill 32BvsDeepSeek R1 Distill 14B

DeepSeek R1 Distill 8BvsQwen 3 8B

7B Class

Llama 3.1 8BvsQwen 2.5 7B

Llama 3.1 8BvsMistral 7B

Llama 3.1 8BvsGemma 2 9B

Llama 3.1 8BvsPhi-4

Qwen 2.5 7BvsMistral 7B

Qwen 2.5 7BvsGemma 2 9B

Qwen 2.5 7BvsPhi-4

Mistral 7BvsGemma 2 9B

Mistral 7BvsPhi-4

Gemma 2 9BvsPhi-4

70B Class

Llama 3.1 70BvsQwen 2.5 72B

Llama 3.1 70BvsLlama 3.3 70B

Llama 3.1 70BvsMistral Large 2

Llama 3.3 70BvsQwen 2.5 72B

Llama 3.3 70BvsMistral Large 2

Qwen 2.5 72BvsMistral Large 2

Large Models

Llama 3.1 70BvsDeepSeek V3

Llama 3.1 405BvsDeepSeek V3

Mixtral 8x22BvsQwen 2.5 72B

Llama 3.1 405BvsQwen 3 235B

Qwen 3 235BvsDeepSeek V3

DeepSeek V3 vs

DeepSeek V3vsQwen 2.5 72B

DeepSeek V3vsMistral Large 2

DeepSeek V3vsLlama 3.3 70B

DeepSeek V3vsQwen 3 235B

DeepSeek V3vsLlama 4 Maverick

DeepSeek V3vsLlama 4 Scout

DeepSeek V3vsQwen 3 32B

DeepSeek V3vsGemma 3 27B

DeepSeek V3vsDeepSeek V3-0324

DeepSeek V3vsMixtral 8x22B

Mid-Size Models

Gemma 3 12BvsPhi-4

Gemma 3 27BvsQwen 3 32B

Phi-4vsQwen 2.5 Coder 14B

Mistral Nemo 12BvsGemma 3 12B

Mistral Small 24BvsGemma 3 27B

Qwen 3 32BvsMistral Small 24B

Qwen 3 32BvsPhi-4

Mistral Nemo 12BvsPhi-4

Small Models

Llama 3.2 3BvsGemma 3 4B

Llama 3.2 1BvsQwen 2.5 1.5B

Gemma 2 2BvsPhi 2

Qwen 3 4BvsLlama 3.2 3B

Gemma 3 1BvsQwen 3 0.6B

Gemma 3 4BvsGemma 3 2B

Code

DeepSeek Coder V2 236BvsCode Llama 34B

StarCoder2 15BvsCode Llama 13B

Qwen 2.5 Coder 7BvsDeepSeek Coder 6.7B

Qwen 2.5 Coder 14BvsStarCoder2 15B

Qwen 2.5 Coder 7BvsCode Llama 7B

Popular

DeepSeek R1vsLlama 3.1 405B

DeepSeek R1vsLlama 4 Maverick

Qwen 3 235BvsLlama 3.1 405B

Qwen 3 32BvsLlama 3.3 70B

MoE vs Dense

Mixtral 8x7BvsLlama 3.1 70B

Mixtral 8x22BvsLlama 3.1 405B

Qwen 3 30B-A3BvsQwen 2.5 7B

DeepSeek V3vsLlama 3.1 405B

Cross-Size

Llama 3.1 8BvsLlama 3.1 70B

Llama 3.1 8BvsLlama 3.1 405B

Llama 3.1 70BvsLlama 3.1 405B

Qwen 2.5 7BvsQwen 2.5 72B

Qwen 3 8BvsQwen 3 32B

Qwen 3 4BvsQwen 3 32B

Gemma 3 4BvsGemma 3 27B

Gemma 3 12BvsGemma 3 27B

DeepSeek R1 Distill 8BvsDeepSeek R1 Distill 70B

Vision

Llama 3.2 11B VisionvsLlama 3.2 90B Vision

Qwen 2.5 VL 7BvsQwen 2.5 VL 72B

Phi 3.5 VisionvsLlama 3.2 11B Vision

Trending

Llama 4 ScoutvsLlama 4 Maverick

Qwen 3 8BvsLlama 3.1 8B

Qwen 3 4BvsGemma 3 4B

Gemma 3 27BvsPhi-4

Qwen 3 8BvsGemma 3 12B

Qwen 3 32BvsGemma 3 27B

Qwen 3 235BvsLlama 4 Maverick

Llama 4 ScoutvsQwen 3 8B

Llama 4 ScoutvsGemma 3 12B

Llama 4 MaverickvsLlama 3.1 405B

Gemma 3 12BvsGemma 2 9B

Qwen 3 8BvsQwen 2.5 7B

DeepSeek V3-0324vsDeepSeek R1

Mistral Small 3.1 24BvsGemma 3 27B

Provider Comparisons

Side-by-side GPU cloud and inference provider comparisons.

RunPodvsLambda Labs

Amazon Web ServicesvsGoogle Cloud Platform

Together AIvsFireworks AI

RunPodvsCoreWeave

Amazon Web ServicesvsMicrosoft Azure

GroqvsFireworks AI

DeepInfravsTogether AI

Lambda LabsvsCoreWeave

Google Cloud PlatformvsMicrosoft Azure

RunPodvsVast.ai

GroqvsTogether AI

DeepInfravsFireworks AI

Amazon Web ServicesvsRunPod

GroqvsDeepInfra

Together AIvsReplicate