GPT-3.5 Turbo (0125): benchmark results
OpenAI's GPT-3.5 Turbo checkpoint gpt-3.5-turbo-0125, released in the API on February 1, 2024. Provider: OpenAI. Released 2024-02-01. Access: API.
Aggregate Rating 1452 ± 8, rank #984 of 1656 rated models, from 650 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| SpeechMap Compliance | 94.5 | % Requests Completed | 97.1 |
| CanItEdit (Lazy) | 42.71 | pass@1 (%; share of the 105 hand-written Python editing prob | 95.8 |
| ConflictQA - KG Conflict (Non-Complementary) | 36.41 | Exact match (%) on 802 single-hop (non-complementary) questi | 90.9 |
| LogicKor - Grammar | 8 | Score (0-10) | 90.5 |
| LogicKor - Understanding | 9.57 | Score (0-10) | 88 |
| Open Portuguese LLM - OAB Exams | 54.31 | Accuracy (%) | 85.8 |
| HELM AIR-Bench 2024 - #6.24: Accounting | 63.33 | Refusal Rate (%) | 84.9 |
| CanItEdit (Descriptive) | 48.14 | pass@1 (%; share of the 105 hand-written Python editing prob | 83.3 |
| Open Portuguese LLM - BLUEX | 62.45 | Accuracy (%) | 82.4 |
| Open Portuguese LLM - ENEM | 72.15 | Accuracy (%) | 82.3 |
| LogicKor - Multi-Turn | 7.88 | Score (0-10) | 80.9 |
| HELM AIR-Bench 2024 - #39-40.47: Classification of individuals - Beliefs | 58.33 | Refusal Rate (%) | 79.1 |
Interactive version: theaggregate.ai/model?slug=gpt-3-5-turbo-0125 · How It Works · Data refreshed daily, snapshot 2026-10-20.