GPT-4-0613

The costliest most powerful OpenAI model.

Model

GPT-4-0613

Details

Developer

OpenAI

License

Private

Model parameters

Pretraining tokens

13T

Release date

March 2023

Supported context length

Price for prompt tokens*

$30/Million tokens

Price for response tokens*

$60/Million tokens

*Note: Data based on 11/14/2023

Model Performance Across Task-Types

Here's how GPT-4-0613 performed across all three task types

Metric

ChainPoll Score

QA without RAG

0.77

QA with RAG

0.76

Long form text generation

0.83

Model Info Across Task-Types

Digging deeper, here’s a look how GPT-4-0613 performed across specific datasets

Tasks	Insights	Dataset Name	Dataset Performance
QA without RAG	The model performs really well which show less bias and good factual knowledge.	Truthful QA	0.67
QA without RAG		Trivia QA	0.87
QA with RAG	The model excels at this task which demonstrates great reasoning and comprehension skills. Compared to other models, it shows better mathematical skills by performing well on DROP.	MS Marco	0.92
		Hotpot QA	0.73
		Drop	0.6
		Narrative QA	0.79
Long form text generation	The model excels at this task which demonstrates great ability to generate long text without factual errors.	Open Assistant	0.83

💰 Cost insights

A very costly model which you can you safely use for any task. It is 30x costlier compared to GPT3.5. One caveat of using it is low quota limits which can be a problem in production. Since it performs closely to GPT3.5 variants, you can consider using it instead.

LLMHALLUCINATIONINDEXLLM