Llama 3.2 90B
Llama 3.2 90BCompare inonMMMU validation
Alternatives?
ModelScoreCostPer taskWhat changes
CheaperLlama 4 MaverickunstatedOpen+22%-58%$0.19 / $0.65
BetterGPT-5.1unstated+42%+377%$1.25 / $10dOpenDots.vlm1.inst+33%——
CheaperQVQ 72B PreviewOpen+17%-57%$0.25 / $0.50CheaperLlama 4 ScoutunstatedOpen+15%-79%$0.10 / $0.30ModelScoreCostPer taskWhat changes
Llama 4 MaverickunstatedOpen+22%-58%$0.19 / $0.65QVQ 72B PreviewOpen+17%-57%$0.25 / $0.50Llama 4 ScoutunstatedOpen+15%-79%$0.10 / $0.30
Gemma 3 27BunstatedOpen+8%-86%$0.08 / $0.16Gemma 3 12BunstatedOpensame-91%$0.05 / $0.10ModelScoreCostPer taskWhat changes
GPT-5.1unstated+42%+377%$1.25 / $10o3unstated+37%+386%$2 / $8
Claude Opus 4.5unstated+34%+1,289%$5 / $25dDots.vlm1.instOpen+33%——Gemini 2.5 Pro Preview (May' 25)unstated+32%——No image inputModelScoreCostPer taskWhat changes
dDots.vlm1.instOpen+33%——iIntern S1Open+29%——
Skywork R1V3 38BOpen+26%——Skywork R1V2 38BOpen+22%——Llama 4 MaverickunstatedOpen+22%-58%$0.19 / $0.65ModelScoreCostPer taskWhat changes
Llama 4 MaverickunstatedOpen+22%—$0.19 / $0.65Llama 4 ScoutunstatedOpen+15%—$0.10 / $0.30Llama 3.2 11B-16%—$0.16 / $0.16Muse Spark OpenxhighOpenn/a—$0.84Reasoning off leaves output unchanged · Self-hosted / llama.cpp+1Muse Wondermaxn/a—$4.52Only auto tool_choice is supported+5Muse Wonder Fulln/a——Only in Muse CodeMuse Spark 1.3xhighn/a—$1.37Only auto tool_choice is supported+5Muse Glimmer 30BhighOpenn/a—$0.058All Meta models →Benchmarks
in Vision & DocumentsMMMU · validation60.3#23
FAQ
How much does Llama 3.2 90B cost?›
$0.72 per 1M input tokens and $0.72 per 1M output tokens (cheapest host: AWS Bedrock).
Is Llama 3.2 90B free to use?›
There is no free API; the cheapest paid price is $0.72 per 1M input tokens.
How good is Llama 3.2 90B, and what is it best at?›
#23 of 41 in Vision & Documents.
Is Llama 3.2 90B better than Nova Pro?›
On MMMU · validation, Llama 3.2 90B scores 60.3 vs 62 — lower.
How can I use Llama 3.2 90B?›
Through the API of AWS Bedrock.
Can Llama 3.2 90B read images?›
It accepts text and images as input.
What is Llama 3.2 90B?›
Llama 3.2 90B is a language model from Meta.
What are the best alternatives to Llama 3.2 90B?›
Llama 4 Maverick (cheaper), GPT-5.1 (better), Dots.vlm1.inst (open weights).