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

Can't find a comparison?

Select any two models from the leaderboard to generate a custom comparison. We cover 262+ models across 60 GPUs.