Skip to main content
Ollama ExplorerBeta
localtools

llama3.1

Llama 3.1 is a new state-of-the-art model from Meta available in 8B, 70B and 405B parameter sizes.

120.4M pulls93 tagsSource updated Nov 30, 2024128K max context

Quick start

ollama run llama3.1:latest

Runnable variants

Memory is a rough file-size estimate; context length, runtime and GPU offload change it.

12 of 93 matching variants shown.

405bTypelocalInputstextDigestdbd6b9ea93deDownload243 GBParameters405bQuantizationUnknownContext128KEst. memory≈ 273 GB
ollama run llama3.1:405b
405b-instruct-fp16TypelocalInputstextDigest8ca13bcda28bDownload812 GBParameters405bQuantizationFP16Context128KEst. memory≈ 908.5 GB
ollama run llama3.1:405b-instruct-fp16
405b-instruct-q2_KTypelocalInputstextDigest57c259aa0162Download149 GBParameters405bQuantizationQ2_KContext128KEst. memory≈ 168 GB
ollama run llama3.1:405b-instruct-q2_K
405b-instruct-q3_K_LTypelocalInputstextDigestfcd77ab1d261Download213 GBParameters405bQuantizationQ3_K_LContext128KEst. memory≈ 239.5 GB
ollama run llama3.1:405b-instruct-q3_K_L
405b-instruct-q3_K_MTypelocalInputstextDigest9ab0ea61e9d6Download195 GBParameters405bQuantizationQ3_K_MContext128KEst. memory≈ 219 GB
ollama run llama3.1:405b-instruct-q3_K_M
405b-instruct-q3_K_STypelocalInputstextDigest5629943379daDownload175 GBParameters405bQuantizationQ3_K_SContext128KEst. memory≈ 197 GB
ollama run llama3.1:405b-instruct-q3_K_S
405b-instruct-q4_0TypelocalInputstextDigest65fa6b82bfdaDownload229 GBParameters405bQuantizationQ4_0Context128KEst. memory≈ 257 GB
ollama run llama3.1:405b-instruct-q4_0
405b-instruct-q4_1TypelocalInputstextDigest283fcc3a816aDownload254 GBParameters405bQuantizationQ4_1Context128KEst. memory≈ 285 GB
ollama run llama3.1:405b-instruct-q4_1
405b-instruct-q4_K_MTypelocalInputstextDigestdbd6b9ea93deDownload243 GBParameters405bQuantizationQ4_K_MContext128KEst. memory≈ 273 GB
ollama run llama3.1:405b-instruct-q4_K_M
405b-instruct-q4_K_STypelocalInputstextDigest4a745087e617Download231 GBParameters405bQuantizationQ4_K_SContext128KEst. memory≈ 259.5 GB
ollama run llama3.1:405b-instruct-q4_K_S
405b-instruct-q5_0TypelocalInputstextDigest441bbd85e453Download279 GBParameters405bQuantizationQ5_0Context128KEst. memory≈ 313 GB
ollama run llama3.1:405b-instruct-q5_0
405b-instruct-q5_1TypelocalInputstextDigestb80f0c0aea72Download305 GBParameters405bQuantizationQ5_1Context128KEst. memory≈ 342 GB
ollama run llama3.1:405b-instruct-q5_1

Source-listed applications

Claude CodeHermes AgentOpenClawOpenCode

Related models