Llama 3.2 11B Vision Instruct
Sep 2024
Llama 3.2 11B Vision InstructCompare inonArena Vision
Alternatives?
ModelScoreCostPer taskWhat changes
CheaperGLM 5.3 FlashunstatedOpenwins 87%-31%$0.15 / $0.50Memory (4-bit): 6.2 GB → 193 GB+5
BetterGPT-5 Minihighwins 80%+99%$0.053
OpenGemma 4 26B A4Bunstatedwins 84%-24%$0.15 / $0.60Memory (4-bit): 6.2 GB → 15 GBCheaperGemma 4 31BunstatedOpenwins 86%-54%$0.10 / $0.34Memory (4-bit): 6.2 GB → 19 GB
CheaperMiMo v2.6 FlashunstatedOpenwins 84%-49%$0.14 / $0.28Memory (4-bit): 6.2 GB → 186 GB+4ModelScoreCostPer taskWhat changes
GLM 5.3 FlashunstatedOpenwins 87%-31%$0.15 / $0.50Memory (4-bit): 6.2 GB → 193 GB+5Gemma 4 31BunstatedOpenwins 86%-54%$0.10 / $0.34Memory (4-bit): 6.2 GB → 19 GBGemma 4 26B A4BunstatedOpenwins 84%-24%$0.15 / $0.60Memory (4-bit): 6.2 GB → 15 GBMiMo v2.6 FlashunstatedOpenwins 84%-49%$0.14 / $0.28Memory (4-bit): 6.2 GB → 186 GB+4MiMo v2.5unstatedOpenwins 83%-49%$0.14 / $0.28Memory (4-bit): 6.2 GB → 186 GBModelScoreCostPer taskWhat changes
GPT-5 Minihighwins 80%+99%$0.053GPT-6 Lunamaxwins 81%-42%$0.068Function calling only with none · Chat Completions+3GPT-5.6 Lunaxhighwins 84%+30%$0.085GPT-5.6 Terraxhighwins 85%+1,204%$0.63Prompts with >272K input tokensGPT-6.1 Solmaxwins 86%+1,059%$0.72Remove unsupported parameters+5ModelScoreCostPer taskWhat changes
Gemma 4 26B A4BunstatedOpenwins 84%-24%$0.15 / $0.60Memory (4-bit): 6.2 GB → 15 GB
Mistral Small 3.1 24BunstatedOpenwins 73%+17%$0.35 / $0.56Context: 131K → 128K+1
Molmo2 8BunstatedOpenwins 66%——Context: 131K → 37K+1AMolmo 7B D 0924Opensame——Context: 131K → 4KOMiniCPM V 2 6Openwins 42%——Context: not verifiedModelScoreCostPer taskWhat changes
Muse Sparkunstatedwins 88%—$1.25 / $4.25Only n=1 · Meta / Chat Completions+1Muse Spark 1.2xhighwins 87%—$0.97Only auto tool_choice is supported+1Muse Spark 1.3xhighwins 87%—$1.37Only auto tool_choice is supported+5Muse Spark 1.1unstatedwins 87%—$1.25 / $4.25Only n=1 · Meta / Chat Completions+1Llama 4 MaverickunstatedOpenwins 73%—$0.19 / $0.65Memory (4-bit): 6.2 GB → 241 GBLlama 4 ScoutunstatedOpenwins 71%—$0.10 / $0.30Memory (4-bit): 6.2 GB → 65 GBLlama 3.2 90B Vision InstructOpenwins 55%—$0.35 / $0.40Context: 131K → 33K+1Muse Spark OpenxhighOpenn/a—$0.84Memory (4-bit): 6.2 GB → 493 GB+2All Meta models →Run it yourself
estimateFits on✓ 8 GB (phone, laptop) · compressed✓ 16 GB · 8-bit✓ 24 GB GPU · full quality✓ 48 GB GPU · full quality✓ 80 GB GPU · full quality
LicenceLlama 3.2 licencecommercial use with conditions
Benchmarks
in Vision & DocumentsArena Vision · overall966.4#84
FAQ
How much does Llama 3.2 11B Vision Instruct cost?›
$0.34 per 1M input tokens and $0.34 per 1M output tokens (cheapest host: DeepInfra).
Is Llama 3.2 11B Vision Instruct free to use?›
There is no free API; the cheapest paid price is $0.34 per 1M input tokens. The weights are open, so you can run it yourself.
How good is Llama 3.2 11B Vision Instruct, and what is it best at?›
#84 of 88 in Vision & Documents.
Is Llama 3.2 11B Vision Instruct better than Qwen 2 VL 7B Instruct?›
On Arena Vision · overall, Llama 3.2 11B Vision Instruct scores 966.4 vs 989.9 — lower.
How can I use Llama 3.2 11B Vision Instruct?›
Through the API of DeepInfra.
What is the context window of Llama 3.2 11B Vision Instruct?›
131K tokens.
When was Llama 3.2 11B Vision Instruct released? Is there a newer version?›
Released 2024-09-18. It is the current version in its line.
Is Llama 3.2 11B Vision Instruct open source?›
Yes, the weights are open (llama3.2 licence).
Can I run Llama 3.2 11B Vision Instruct locally?›
The weights are open and can be self-hosted; hardware needs are not confirmed yet.
Can Llama 3.2 11B Vision Instruct read images?›
It accepts text and images as input.
What is Llama 3.2 11B Vision Instruct?›
Llama 3.2 11B Vision Instruct is an open-weights language model from Meta, released Sep 2024.
What are the best alternatives to Llama 3.2 11B Vision Instruct?›
GLM 5.3 Flash (cheaper), GPT-5 Mini (better), Gemma 4 26B A4B (open weights).