Meta 🇺🇸 · Llama 3

Llama 3 8B Instruct

Meta's eight-billion-parameter instruction-tuned Llama 3 checkpoint, released for local and broadly hosted assistant deployment with accompanying safety tooling.

DOOM SCORE51.5out of 100model risk profile, not the overall index
0 comments · 0 votesOpen discussion

Public discussion is readable by everyone. Sign in to comment, reply, or vote.

No comments yet. Start the discussion.

CURRENT ASSESSMENT · REVISION 3

Why this model scores 51.5

Instruction tuning increased practical capability and harmful usability, while small size, downloadable weights and many distribution channels produced exceptional reach and weak centralized control.

Capability50
Autonomy10
Deployment97
Misuse potential50
Control difficulty62
MODEL-ATTRIBUTED EVIDENCE

News tied to Llama 3 8B Instruct

The model score of 51.5 rates this model's risk profile. The overall Doom Index of 67.9 measures the complete temporally weighted evidence record. These values answer different questions.

NET MODEL-ATTRIBUTED INDEX CONTRIBUTION+0.05

Each article's current 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 use a bounded temporal offset around their technical profile, but never feed back into the overall index.

Race TOWARD

Meta releases Llama 3 8B and 70B base and instruction models

Meta released Llama 3 in 8-billion and 70-billion parameter base and instruction-tuned variants, with downloadable weights, broad planned cloud distribution and new Llama Guard 2, Code Shield and CyberSec Eval 2 safety tools.

Full item contribution
+0.18
Llama 3 8B Instruct equal share
+0.05
Read assessment →
AUDIT TRAIL

Model score history

  1. R3
    Doom Score 51.5

    Exact-version evidence chronology replayed after run doombench-hourly-news-20260812-142416 under temporal-monthly-pressure-v4.

    12 Aug 2026
  2. R2
    Doom Score 52.0

    Exact-version evidence chronology replayed after run doombench-hourly-news-20260812-132112 under fixed-sensitivity-v3.

    12 Aug 2026
  3. R1
    Doom Score 53.2

    New exact Llama 3 8B Instruct public model profile.

    12 Aug 2026
SHARE THE FINDINGS

Share this page

DoomBench social sharing card for Llama 3 8B Instruct.
  1. Llama 3 8B Instruct by Meta has a DoomBench model risk score of 51.5 out of 100, based on five transparent version-specific dimensions rather than the overall index.

  2. Llama 3 8B Instruct's highest current DoomBench dimension is deployment at 97.0 out of 100; the profile publishes every component score and its editorial rationale.

  3. DoomBench links 1 source-backed evidence item to Llama 3 8B Instruct, while keeping the model's risk profile separate from each item's contribution to the live Doom Index.

    https://www.doombench.com/models/meta-llama-3-8b-instruct