Meta · Code Llama

Code Llama 70B

READER SUMMARY

Open-weight 70B base checkpoint specialized for code generation and completion.

DOOM SCORE50.5out of 100model risk profile, not the overall index

CURRENT ASSESSMENT · REVISION 1

Why this model scores 50.5

A large code-specialized base model offered meaningful programming capability with highly diffuse open weights, while lacking instruction tuning or intrinsic autonomous behavior.

Capability51

Autonomy4

Deployment94

Misuse potential45

Control difficulty58

0 comments · 0 votesOpen discussion

No comments yet. Start the discussion.

MODEL-ATTRIBUTED EVIDENCE

News tied to Code Llama 70B

The model score of 50.5 rates this model's risk profile. The overall Doom Index of 61.2 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.

29 Jan 2024RaceTOWARD

Meta releases three open-weight Code Llama 70B variants

Meta released Code Llama 70B base, Python-specialized and instruction-tuned checkpoints under the Llama 2 community license, materially expanding openly deployable code-generation capability.

Full item contribution: +0.04
Code Llama 70B equal share: +0.01

Read assessment →

AUDIT TRAIL

Model score history

R1
Doom Score 50.5
New exact Code Llama 70B base profile.
12 Aug 2026