89.4%
HELM Safety
Models / OpenAI
GPT-4o (2024-05)
GPT-4o · Released 2024-05-13
OpenAI's May 2024 GPT-4o flagship. MMLU 88.7% and HELM safety 89.4% on public leaderboards. The reference GPT-4o checkpoint before GPT-5.
Legacy OpenAI multimodal baseline — use GPT-5.4/5.5 for new work.
Composite89
Context128K
Input / 1M$2.5
Output / 1M$10
Knowledge cutoff2023-10-01
Statuslightweight
Not enough families covered yet.Add more benchmark rows for GPT-4o (2024-05) to see the radar.
Benchmark placements
Per-benchmark positions
Where GPT-4o (2024-05) places on each public benchmark source that publishes a row. Rank counts every model with a latest row in the same test — not a universal quality score.
| Benchmark | Family | Rank | Score | Source | Date |
|---|---|---|---|---|---|
| HELM Safety | Safety | #4/ 9 | 89 | Stanford HELM leaderboard | 2024-08-15 |
Family context
Strongest rows in each family
The three highest-scoring carded models in each capability family. Where GPT-4o (2024-05) shows up, it's highlighted.
Knowledge
- 191
- 291
- 391
Reasoning
- 195
- 294
- 394
Math
- 199
- 297
- 396
Coding
- 196
- 296
- 396
Agentic
- 1100
- 2100
- 3100
Multimodal
- 172
- 272
- 371
Long context
- 175
- 275
- 374
Tool use
- 189
- 287
- 387
SafetyGPT-4o (2024-05) · 89
- 192
- 292
- 391
Human preference
- 195
- 292
- 391
Family coverage
Best score per family
A high composite that hides a weak family is a trap. These bars surface the families where this model hasn't been publicly tested, and where it leads.
Price vs. performance
Where GPT-4o (2024-05) sits on the scatter
Hover or tab any dot for name, score, and input price.
AnthropicOpenAIMistralAlibabaDeepSeekGoogleMoonshotMiniMaxxAIMetaZhipuXiaomiMeituanCohere
Full benchmark ledger
Every catalog benchmark for GPT-4o (2024-05). Scores link to the original source; gaps mean no public row exists yet.
1 of 32 catalog benchmarks have a sourced row for GPT-4o (2024-05).
Safety
Sources
Where the numbers come from
- GPT-4o announcement· checked 2026-06-09