OpenAIproprietary
Released: 2026-06-01

OpenAI: GPT-4

OpenAI's flagship model, GPT-4 is a large-scale multimodal language model capable of solving difficult problems with greater accuracy than previous models due to its broader general knowledge and advanced reasoning...

Intelligence Index
91.5/ 100
Calibrated multi-domain composite
Throughput Speed
55tok/sec
Output streaming throughput
Time To First Token
420ms
Initial chunk server latency
LiveBench Contam-Free
80.5%
September 2026 suite score

LiveBench Multi-Domain Evaluation Breakdown

Monthly refreshed contamination-free test sets
Logical Reasoning84.2%
Coding & Repo Repair82.4%
Mathematics (AIME)81.4%
Data Analysis & Tables77.8%
Language & Comprehension79.6%
Instruction Following80.5%

Sourced Benchmark Evaluations (51)

Verified performance across authoritative benchmarks with provenance tracking.
BenchmarkStatusScoreTrustDateSource TypeProvenance
ARC-AGI-3Active57.7%
93
2026-07-06independentSource ↗
ARC-AGI-2Nearing Saturation69.3%
75
2026-06-29independentSource ↗
FrontierMathActive23.8%
98
2026-06-26independentSource ↗
PutnamBenchActive44.3%
90
2026-06-25independentSource ↗
Omni-MATHDeprecated63%
10
2026-06-23independentSource ↗
SWE-LancerDeprecated59.1%
20
2026-06-23independentSource ↗
SoohakActive64.7%
98
2026-06-21independentSource ↗
FrontierCodeActive50.4%
98
2026-06-21independentSource ↗
AIMENearing Saturation80.8%
75
2026-06-19independentSource ↗
Humanity's Last ExamActive39.1%
93
2026-06-19independentSource ↗
StructFlowBenchNearing Saturation68.4%
75
2026-06-19independentSource ↗
OlympiadBenchActive69.8%
80
2026-06-18independentSource ↗
MathArenaActive68.1%
99
2026-06-17independentSource ↗
FireBenchActive67.6%
95
2026-06-17independentSource ↗
Terminal-BenchActive60.8%
80
2026-06-17independentSource ↗
CodeContestsDeprecated56.5%
10
2026-06-16independentSource ↗
SimpleQAActive47.6%
90
2026-06-16independentSource ↗
SWE-bench ProActive53.9%
90
2026-06-15independentSource ↗
ComplexBenchNearing Saturation66%
80
2026-06-15independentSource ↗
MultiPL-EDeprecated79.9%
10
2026-06-14independentSource ↗
AgentIFNearing Saturation54.7%
80
2026-06-13independentSource ↗
MATHSaturated83.4%
25
2026-06-13independentSource ↗
LiveCodeBenchSaturated71.3%
39
2026-06-13independentSource ↗
MercuryDeprecated76.5%
10
2026-06-12independentSource ↗
MMLU-ProNearing Saturation84.2%
75
2026-06-12independentSource ↗
CyberSecEvalDeprecated78.2%
10
2026-06-11independentSource ↗
TabMWPSaturated78.3%
25
2026-06-11independentSource ↗
SWE-bench VerifiedSaturated67.8%
25
2026-06-10independentSource ↗
MMLU-ReduxSaturated89.7%
25
2026-06-10independentSource ↗
SWE-benchDeprecated59.1%
10
2026-06-09independentSource ↗
MGSMNearing Saturation80.8%
65
2026-06-09independentSource ↗
TruthfulQASaturated80.5%
25
2026-06-08independentSource ↗
AGIEvalSaturated76.4%
25
2026-06-08independentSource ↗
SuperGLUESaturated90.6%
25
2026-06-07independentSource ↗
GSM-HardDeprecated81.7%
10
2026-06-07independentSource ↗
MT-BenchNearing Saturation8.3%
75
2026-06-06independentSource ↗
CodeXGLUEDeprecated81.7%
23
2026-06-06independentSource ↗
BIG-Bench HardSaturated87%
25
2026-06-06independentSource ↗
MBPP+Nearing Saturation88.6%
65
2026-06-05independentSource ↗
IFEvalDeprecated78.9%
25
2026-06-05independentSource ↗
WinoGrandeSaturated92.4%
28
2026-06-05independentSource ↗
GPQANearing Saturation80.5%
78
2026-06-05independentSource ↗
HumanEval+Saturated93.9%
25
2026-06-04independentSource ↗
CommonsenseQASaturated93.2%
28
2026-06-04independentSource ↗
MBPPSaturated92.1%
25
2026-06-03independentSource ↗
GSM8KSaturated91.9%
25
2026-06-03independentSource ↗
GLUESaturated95.9%
28
2026-06-03independentSource ↗
MMLUSaturated93.3%
25
2026-06-02independentSource ↗
ARC-AGISaturated83.5%
25
2026-06-02independentSource ↗
HumanEvalSaturated99.1%
25
2026-06-02independentSource ↗
LiveBench (Continuous)Active80.5%
94
2025-02-01independentSource ↗
OpenAI Compatible API ExecutionAPI Ready
from openai import OpenAI

client = OpenAI()
response = client.chat.completions.create(
    model="openai-gpt-4",
    messages=[{"role": "user", "content": "Evaluate multi-step logic problem"}],
    temperature=0.2
)
print(response.choices[0].message.content)

Evaluation & Sourcing Notes

Scores listed for OpenAI: GPT-4 represent verified evaluations extracted from official research papers, independent evaluation suites (HELM, LMSYS, OpenCompass, LiveBench), and verified audit reports.

All benchmarks marked as saturated or deprecated reflect historical performance where the benchmark no longer provides active discriminative power.