GPT-4 Benchmark Details

GPT-4 currently shows benchmark results led by GSM8K (11 / 70, score 92), MMLU (32 / 124, score 86.40), HumanEval (39 / 140, score 82). This page also compares it with 1 competitor models and 2 predecessor or same-series models, including performance and pricing views when available. 1 source link is attached for reference.

Benchmark Results

GPT-4

Benchmark Results

General Knowledge

3 evaluations

Benchmark / mode

Score

Rank/total

MMLU

Standard Mode

86.40

32 / 124

MMLU

Standard Mode

86.40

32 / 124

C-Eval

Standard Mode

68.70

21 / 48

Math and Reasoning

1 evaluations

Benchmark / mode

Score

Rank/total

GSM8K

Standard Mode

11 / 70

Coding and Software Engineer

2 evaluations

Benchmark / mode

Score

Rank/total

HumanEval

Standard Mode

39 / 140

HumanEval

Standard Mode

63 / 140

Reading Comprehension

1 evaluations

Benchmark / mode

Score

Rank/total

DROP

Standard Mode

80.90

7 / 9

Writing and Creative Capabilities

1 evaluations

Benchmark / mode

Score

Rank/total

Creative Writing

Standard Mode

749.90

99 / 106

Compare with other models

Competitor Comparison

Benchmark scores for GPT-4 compared against top models in its class

GPT-4Claude3-Opus

Benchmark categories:

The chart shows each model’s highest score per benchmark within the current filter. Out-of-100 benchmarks use raw heights; out-of-range benchmarks are scaled within that benchmark while labels keep the original scores.

4 benchmarks with comparable scores. Each model shows its best score; mode label is displayed below.

Benchmark	GPT-4Current	Claude3-Opus
MMLU Accuracy 综合评估	86.40Standard Mode	86.80Standard Mode
GSM8K Accuracy 数学推理	92.00Standard Mode	95.00Standard Mode
HumanEval Accuracy 编程与软件工程	82.00Standard Mode	84.90Standard Mode
DROP F1 阅读理解	80.90Standard Mode	83.10Standard Mode

Standard API Pricing: GPT-4 vs. Peer Models

Shows standard text input and output pricing side by side for each model. If extended-context pricing exists, the chart keeps the base rate and explains the threshold below.

Source: DataLearnerAI. Standard text prices shown here use the default supplier. · USD / 1M tokens

Model	Supplier	Standard input	Standard output	Base price applies to
Claude3-Opus	Anthropic	$15 / 1M tokens	$75 / 1M tokens	—

Version History

How each version of the GPT-4 series stacks up on benchmark tests

GPT-4GPT-3.5 GPT-3

Benchmark categories:

4 benchmarks with comparable scores. Each model shows its best score; mode label is displayed below.· Click a row to view its trend chart.

Benchmark	GPT-4Current	GPT-3.5	GPT-3
C-Eval Accuracy 综合评估	68.70Standard Mode	54.40Standard Mode	--
MMLU Accuracy 综合评估	86.40Standard Mode	70.00Standard Mode	53.90Standard Mode
GSM8K Accuracy 数学推理	92.00Standard Mode	57.10Standard Mode	--
HumanEval Accuracy 编程与软件工程	82.00Standard Mode	48.10Standard Mode	--

Single-Benchmark Version Trend

Viewing: C-Eval · 综合评估

Benchmark

NormalNormal + ToolsThinkingThinking + ToolsDeepDeep + Tools

X-axis shows model and release date, Y-axis shows score; solid lines connect the same mode across versions, while dotted guides align modes within the same generation.

Standard API Pricing Across the GPT-4 Series

Shows standard text input and output pricing side by side for each model. If extended-context pricing exists, the chart keeps the base rate and explains the threshold below.

Source: DataLearnerAI. Standard text prices shown here use the default supplier.

Comparable standard text pricing is not available for these models.

Sources

openai.comopenai.com