Model

GPT-3 175B

CLOSED

OpenAI · GPT-3 family · released Jun 11, 2020

175B few-shot learner that launched the OpenAI API in private beta; base GPT-3 models were retired from the API on 2024-01-04.

ReasoningCodingVisionFunction callingTool useAgentic

197.2

Elo · rank #363

Parameters

175B

Active params

175B (dense)

Context

2K tokens

Architecture

Dense decoder-only Transformer

License

Proprietary

Languages

—

API price (in/out)

No hosted API

Modalities

text

Benchmark results

Bar shows position within the tracked field; marker = field best

ANLIReasoning36.8%#8

best: Phi-3-small (7B) · 58.1%

ARC-ChallengeReasoning51.5%#119

best: Llama 3.1 405B · 96.9%

ARC-EasyReasoning70.1%#49

best: Phi-3-medium (14B) · 97.7%

DROPReasoning36.5#58

best: Hunyuan-T1 · 93.1

HellaSwagReasoning79.3%#86

best: Claude 3 Opus · 95.4%

LAMBADAReasoning86.4%#1

best: this model · 86.4%

MMLUKnowledge43.9%#252

best: OpenAI o3 · 92.9%

OpenBookQAReasoning65.4%#10

best: Claude 1 · 90.8%

PIQAReasoning82.8%#19

best: GPT-4o mini · 93.1%

RACE-HReasoning46.8%#17

best: Claude 3 Opus · 92.9%

StoryClozeReasoning87.7%#1

best: this model · 87.7%

SuperGLUEReasoning71.8%#4

best: UL2 20B · 90.7%

TriviaQAKnowledge71.2%#23

best: Sarvam-1 (2B) · 90.6%

TruthfulQAKnowledge58.0%#29

best: Phi-3.5-MoE (16x3.8B, 6.6B active) · 77.5%

WinoGrandeReasoning77.7%#46

best: PaLM 2 · 90.9%

Run it locally

Closed weights — available via API only. No local deployment.

Input / M tok

—

Output / M tok

—

GPT-3 family

Elo progression across releases

GPT-3 175BJun 2020197.2

Resources

GHopenai/gpt-3

API price — · each benchmark row carries its own source badge (see methodology)