Meta Llama 3 8B Instruct
Apr 2024
Meta Llama 3 8B InstructCompare inon
Meta Llama 3 8B Instruct measured on 3 of 104 Chat & Knowledge benchmarks3
BenchmarkByScoreRank
Legacy · saturated or no longer reported2
No benchmark matches
Alternatives?
ModelScoreCostPer taskWhat changes
CheaperLlama 3.1 8B InstructunstatedOpensame-23%$0.02 / $0.04
BetterGPT-5.6 Lunalow×34.3+1,285%$0.01
OpenQwen3 14Bunstated×34.1+1,785%$0.35 / $1.40Memory (4-bit): 4 GB → 7.4 GBCheaperLlama 3.2 1B InstructunstatedOpen-71%-81%$0.005 / $0.01No strict JSONBetterGPT OSS 20BhighOpen×26.1+11%$0.011Memory (4-bit): 4 GB → 13 GBModelScoreCostPer taskWhat changes
Llama 3.1 8B InstructunstatedOpensame-23%$0.02 / $0.04Llama 3.2 1B InstructunstatedOpen-71%-81%$0.005 / $0.01No strict JSONModelScoreCostPer taskWhat changes
GPT-5.6 Lunalow×34.3+1,285%$0.01GPT OSS 20BhighOpen×26.1+11%$0.011Memory (4-bit): 4 GB → 13 GB
Gemini 3.1 Flash Litehigh×41.1+1,631%$0.06GPT-6 Lunamax×50.9+515%$0.068Function calling only with none · Chat Completions+3
Claude Haiku 5.5high×50+515%$0.079Remove sampling parameters+5ModelScoreCostPer taskWhat changes
Qwen3 14BunstatedOpen×34.1+1,785%$0.35 / $1.40Memory (4-bit): 4 GB → 7.4 GBQwen 3.5 9BnoneOpen×31.7+185%$0.08 / $0.13Memory (4-bit): 4 GB → 4.8 GBQwen3 8BunstatedOpen×28.9+854%$0.18 / $0.70Memory (4-bit): 4 GB → 4.1 GBQwen3.5 4BnoneOpen×28.7+85%$0.03 / $0.15Qwen3 4B 2507 InstructunstatedOpen×26.9+1,438%$0.40 / $0.80ModelScoreCostPer taskWhat changes
Muse Wondermax×51.1—$4 / $20Only auto tool_choice is supported+5Muse Spark 1.3xhigh×51—$1.37Only auto tool_choice is supported+5Muse Spark Open 816B A42B×50.4—$1.25 / $4.25Context: not verifiedMuse Sparkunstated×45.7—$1.25 / $4.25Only n=1 · Meta / Chat Completions+1Llama 4 Maverick 17B 128E Instruct FP8Open×10.6—$0.20 / $0.80Memory (4-bit): 4 GB → 241 GBLlama 3.1 405B Instruct×5——Llama 4 ScoutunstatedOpen×4—$0.10 / $0.30Memory (4-bit): 4 GB → 65 GBLlama 3.3 70B InstructunstatedOpen×2.6—$0.10 / $0.32Memory (4-bit): 4 GB → 42 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
in Chat & KnowledgeGPQA Diamond · Epoch26.1#126
MATH Level 56.1#33
OTIS Mock AIME 2024 20251.9#121
FAQ
How much does Meta Llama 3 8B Instruct cost?›
$0.03 per 1M input tokens and $0.04 per 1M output tokens (cheapest host: DeepInfra).
Is Meta Llama 3 8B Instruct free to use?›
There is no free API; the cheapest paid price is $0.03 per 1M input tokens. The weights are open, so you can run it yourself.
How good is Meta Llama 3 8B Instruct, and what is it best at?›
#126 of 133 in Chat & Knowledge.
Is Meta Llama 3 8B Instruct better than Qwen2.5 7B Instruct?›
On OTIS Mock AIME 2024 2025, Meta Llama 3 8B Instruct scores 1.9 vs 2.5 — about the same.
How can I use Meta Llama 3 8B Instruct?›
Through the API of DeepInfra.
What is the context window of Meta Llama 3 8B Instruct?›
8K tokens.
When was Meta Llama 3 8B Instruct released? Is there a newer version?›
Released 2024-04-17. It is the current version in its line.
Is Meta Llama 3 8B Instruct open source?›
Yes, the weights are open (llama3 licence).
Can I run Meta Llama 3 8B Instruct locally?›
The weights are open and can be self-hosted; hardware needs are not confirmed yet.
Can Meta Llama 3 8B Instruct read images?›
No, it accepts text only.
What is Meta Llama 3 8B Instruct?›
Meta Llama 3 8B Instruct is an open-weights language model from Meta, released Apr 2024.
What are the best alternatives to Meta Llama 3 8B Instruct?›
Llama 3.1 8B Instruct (cheaper), GPT-5.6 Luna (better), Qwen3 14B (open weights).