Sunk Cost sunkcost.ai Data checked 2026-09-03

Llama 3.1 8B Instruct vs Qwen3 32B

Neither has a clear lead on the index. Llama 3.1 8B Instruct is the smaller download at 4.9 GB, so it runs on cheaper hardware.

Llama 3.1 8B InstructQwen3 32B
Intelligence index77
ClassBelow every hosted tierBelow every hosted tier
Weights4.9 GB20 GB
QuantisationQ4_K_MQ4_K_M
Parameters8B32.8B
Max context128k40k
API price per 1M$0.02 in / $0.04 out$0.08 in / $0.28 out
LicenceLlama 3.1 Community LicenseApache 2.0
Cheapest machine that runs itMac mini M6, 16GB $899Strix Halo Framework Desktop, 64GB $1,959
Summarising good good
Translation usable good
Everyday coding usable good
Reasoning & maths don’t usable
Agentic work don’t usable

Ratings are coarse on purpose. Speeds and pay-back depend on the machine — open either model's page for the full list, or see both against the frontier. Context is 32k throughout.