MMLU
57-subject multiple-choice exam spanning STEM, humanities, and social sciences — the defining knowledge benchmark of the 2020–2024 era.
Frontier models cluster at 86–92% since 2024; ~2–6% label errors put the effective ceiling near 90%, so top-model gaps are within noise.
Performance Timeline
Longitudinal progression of model scores against human baselines.Performance & Historical Trajectory
Empirical score progression across model release dates and evaluation rounds.
| Model | Score | Date | Source Type | Provenance |
|---|---|---|---|---|
| Llama 3.1 405B | 88.6% | 2024-07-23 | vendor-reported | Source ↗ |
| GPT-4o | 88.7% | 2024-05-13 | vendor-reported | Source ↗ |
| Claude 3 Opus | 86.8% | 2024-03-04 | vendor-reported | Source ↗ |
| GPT-4 | 86.4% | 2023-03-14 | vendor-reported | Source ↗ |
| PaLM 540B | 69.3% | 2022-04-05 | vendor-reported | Source ↗ |
| Chinchilla 70B | 67.6% | 2022-03-29 | vendor-reported | Source ↗ |
| Gopher 280B | 60% | 2021-12-08 | vendor-reported | Source ↗ |
| GPT-3 175B | 43.9% | 2020-09-07 | independent | Source ↗ |
Human Baseline & Difficulty Horizon
Calibrated human reference points, specialist benchmarks, and ceiling thresholds.Measured human expert accuracy across 57 humanities, STEM, and social science subjects.
Metric & Scoring Methodology
Verification protocols, aggregation formulas, and specialized metric variants.accuracy (%)Dataset & Compute Cost
Evaluation volume, public availability, API pricing, and local hardware requirements.$5 – $20 USD for full benchmark evaluation run on frontier APIs.
1x NVIDIA RTX 4090 (24GB) or A100 (40GB/80GB) via vLLM / SGLang
How to Run & Reproduce
Standardized evaluation protocols, CLI commands, and reproducible runner templates.lm_eval --model hf --model_args pretrained=<model_path> --tasks mmlu --batch_size autoopencompass --datasets mmlu --models <model_config># Standard API Evaluation Loop
from openai import OpenAI
client = OpenAI()
response = client.chat.completions.create(
model="gpt-4o",
messages=[{"role": "user", "content": prompt}],
temperature=0.0,
)When publishing results for MMLU, always report the exact prompt template, few-shot exemplar ordering, sampling temperature (temperature=0), maximum reasoning budget tokens, and the precise timestamped model snapshot ID.
Contamination & Memorization Analysis
Audit of pretraining exposure risks, memorization vectors, and refresh policies.Static fixed snapshot
Public on web / HuggingFace