OpenAI · GPT-3

GPT-3 175B

READER SUMMARY

OpenAI's 175-billion-parameter autoregressive language model, demonstrating broad zero-shot, one-shot, and few-shot task adaptation and later deployed through a controlled API.

DOOM SCORE43.2out of 100model risk profile, not the overall index
CURRENT ASSESSMENT · REVISION 2

Why this model scores 43.2

GPT-3 was a major historical jump in broad language capability and adaptable text generation, with practical API deployment and meaningful misinformation, bias, and scalable-content risks. Centralized private-beta access and limited independent action constrained autonomy and control difficulty.

Capability48
Autonomy10
Deployment45
Misuse potential42
Control difficulty32
MODEL-ATTRIBUTED EVIDENCE

News tied to GPT-3 175B

The model score of 43.2 rates this model's risk profile. The overall Doom Index of 60.7 measures the recalibrated complete evidence corpus. These values answer different questions.

NET MODEL-ATTRIBUTED INDEX CONTRIBUTION+0.06

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.

RaceTOWARD

Microsoft obtains an exclusive license to OpenAI's GPT-3

Microsoft announced an exclusive license to OpenAI's 175-billion-parameter GPT-3 language model, while OpenAI continued offering its public API. The agreement concentrated privileged commercialization rights and deepened the companies' frontier-model partnership.

Full item contribution
+0.02
GPT-3 175B equal share
+0.02
Read assessment →
CapabilityTOWARD

GPT-3 demonstrates broad few-shot learning at 175 billion parameters

OpenAI's original GPT-3 paper reported a 175-billion-parameter autoregressive language model that performed many tasks from instructions or a few examples without gradient updates, while also documenting weaknesses, bias, and human difficulty distinguishing some generated news. OpenAI separately deployed GPT-3-family weights through a controlled private-beta API on June 11.

Full item contribution
+0.04
GPT-3 175B equal share
+0.04
Read assessment →
AUDIT TRAIL

Model score history

  1. R2
    Doom Score 43.2

    Full-corpus evidence recalculated after run intensive-backfill-20260811-211330 under bounded-corpus-v2.

    11 Aug 2026
  2. R1
    Doom Score 40.2

    New exact source-verified historical frontier model absent from the durable catalogue.

    11 Aug 2026