Meta · Llama 3.1

Llama 3.1 70B Instruct

READER SUMMARY

Meta's downloadable 70B instruction-tuned Llama 3.1 model with 128K context, multilingual dialogue and tool use.

DOOM SCORE65.1out of 100model risk profile, not the overall index

CURRENT ASSESSMENT · REVISION 1

Why this model scores 65.1

High general capability, tool use and strong instruction following combine with broad partner hosting and downloadable weights to create substantial misuse potential and weak centralized containment.

Capability71

Autonomy18

Deployment96

Misuse potential68

Control difficulty73

0 comments · 0 votesOpen discussion

No comments yet. Start the discussion.

MODEL-ATTRIBUTED EVIDENCE

News tied to Llama 3.1 70B Instruct

The model score of 65.1 rates this model's risk profile. The overall Doom Index of 63.2 measures the recalibrated complete evidence corpus. These values answer different questions.

NET MODEL-ATTRIBUTED INDEX CONTRIBUTION+0.01

Each article's leave-one-out Doom Index contribution is divided equally among the exact models named on that article. This prevents multi-model evidence from being claimed in full on several model pages. Model risk scores are recalculated from their technical profile and related news, but never feed back into the overall index.

23 Jul 2024CapabilityTOWARD

Meta releases frontier-scale Llama 3.1 open models

Meta released Llama 3.1 in 8B, 70B and 405B base and instruction tiers with 128K context, multilingual support, tool use, downloadable weights and day-one distribution through more than 25 partners.

Full item contribution: +0.05
Llama 3.1 70B Instruct equal share: +0.01

Read assessment →

AUDIT TRAIL

Model score history

R1
Doom Score 65.1
New exact Llama 3.1 70B instruction checkpoint.
12 Aug 2026