Llama 3 8B Instruct
Apr 2024
Llama 3 8B InstructCompare inon
Benchmarks of Llama 3 8B Instruct16
BenchmarkByScoreRank
Legacy · saturated or no longer reported1
No benchmark matches
—per task?
API price not publicprice
—output speed?
8Kcontext
Ranking?
Alternatives?
ModelScoreCostPer taskWhat changes
BetterGPT-5 Minihighwins 68%—$0.053
OpenGemma 3 12Bunstatedwins 69%—$0.05 / $0.10Memory (4-bit): 4 GB → 6.1 GBBetterGPT-6 Lunamaxwins 77%—$0.068Function calling only with none · Chat Completions+3BetterGPT-5.6 Lunaxhighwins 77%—$0.085BetterGPT-5.6 Terraxhighwins 79%—$0.63Prompts with >272K input tokensModelScoreCostPer taskWhat changes
Nothing here yet.
ModelScoreCostPer taskWhat changes
GPT-5 Minihighwins 68%—$0.053GPT-6 Lunamaxwins 77%—$0.068Function calling only with none · Chat Completions+3GPT-5.6 Lunaxhighwins 77%—$0.085GPT-5.6 Terraxhighwins 79%—$0.63Prompts with >272K input tokensGPT-6.1 Solmaxwins 82%—$0.72Remove unsupported parameters+5ModelScoreCostPer taskWhat changes
Gemma 3 12BunstatedOpenwins 69%—$0.05 / $0.10Memory (4-bit): 4 GB → 6.1 GBGemma 3n E4B InstructunstatedOpenwins 64%——
NVIDIA Nemotron 3.5 Lightning 30B A3B NVFP4Openwins 62%——Memory (4-bit): 4 GB → 11 GBGemma 3 4BunstatedOpenwins 61%—$0.04 / $0.05
Granite 4.1 8BunstatedOpenwins 60%——Memory (4-bit): 4 GB → 4.4 GBModelScoreCostPer taskWhat changes
Muse Wondermaxwins 84%—$4.52Only auto tool_choice is supported+5Muse Sparkunstatedwins 82%—$1.25 / $4.25Only n=1 · Meta / Chat Completions+1Muse Spark 1.3maxwins 82%—$1.60Only auto tool_choice is supported+5Muse Spark 1.2xhighwins 81%—$0.97Only auto tool_choice is supported+1Muse Spark 1.1unstatedwins 81%—$1.25 / $4.25Only n=1 · Meta / Chat Completions+1Muse Glimmer 30BunstatedOpenwins 72%—$0.30 / $1.10Memory (4-bit): 4 GB → 18 GBLlama 4 MaverickunstatedOpenwins 65%—$0.19 / $0.65Memory (4-bit): 4 GB → 241 GBLlama 4 ScoutunstatedOpenwins 63%—$0.10 / $0.30Memory (4-bit): 4 GB → 65 GBAll 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 licencecommercial use with conditions
Benchmarks
all categoriesChat & Knowledge
Arena Text · overall1,224#180
Arena Text · Chinese1,076#171
Arena Text · English1,213.2#169
Arena Text · French1,158.8#142
Arena Text · German1,104.4#145
Arena Text · instruction following1,127#174
Arena Text · Japanese967.4#133
Arena Text · Korean1,003.6#138
Arena Text · non-English1,097.5#175
Arena Text · Polish1,137#132
Arena Text · Russian1,109.2#176
Arena Text · Spanish1,172.7#143
Arena Text · overall · full cohort1,165.6#172
Writing & Language
Arena Text · creative writing1,196#177
Arena Text · creative writing · full cohort1,149.7#171
FAQ
How much does Llama 3 8B Instruct cost?›
There is no per-unit API price: API price not public.
Is Llama 3 8B Instruct free to use?›
There is no free API. The weights are open, so you can run it yourself.
How good is Llama 3 8B Instruct, and what is it best at?›
#180 of 196 in Chat & Knowledge, #177 of 195 in Writing & Language.
Is Llama 3 8B Instruct better than Granite 4.2 8B?›
On Arena Text · creative writing, Llama 3 8B Instruct scores 1,196 vs 1,201 — about the same.
What is the context window of Llama 3 8B Instruct?›
8K tokens.
When was Llama 3 8B Instruct released? Is there a newer version?›
Released 2024-04-17. It is the current version in its line.
Is Llama 3 8B Instruct open source?›
Yes, the weights are open (llama3 licence).
Can I run Llama 3 8B Instruct locally?›
The weights are open and can be self-hosted; hardware needs are not confirmed yet.
Can Llama 3 8B Instruct read images?›
No, it accepts text only.
What is Llama 3 8B Instruct?›
Llama 3 8B Instruct is an open-weights language model from Meta, released Apr 2024.
What are the best alternatives to Llama 3 8B Instruct?›
GPT-5 Mini (better), Gemma 3 12B (open weights).