GPT-3.5 Turbo (0301): benchmark results
Provider: OpenAI. Released 2023-03-01. Access: API.
Aggregate Rating 1446 ± 21, rank #3949 of 10916 rated models, from 415 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| ConvRe - Re2Text Hard | 59 | Accuracy (%) | 100 |
| HELM AIR-Bench 2024 - #6.25: Financial | 50 | Refusal Rate (%) | 100 |
| HELM Classic - MATH | 48.83 | Equivalent (%) | 100 |
| HELM Classic - RAFT | 76.82 | Exact Match (%) | 100 |
| HELM AIR-Bench 2024 - #6: Advice in Heavily Regulated Industries | 31.33 | Refusal Rate (%) | 98.8 |
| HELM Classic - Entity Matching | 94.26 | Exact Match (%) | 98.5 |
| HELM Classic - LSAT | 25.22 | Exact Match (%) | 98.5 |
| HELM Classic - MATH Chain-of-Thought | 68.94 | Equivalent (%) | 98.5 |
| HELM Classic - MMLU | 58.98 | Exact Match (%) | 98.5 |
| HELM Classic - QuAC | 51.2 | F1 (%) | 98.5 |
| HELM AIR-Bench 2024 - #4.3: Adversely affects legal rights | 90 | Refusal Rate (%) | 98.3 |
| HELM AIR-Bench 2024 - #5.11: Heavy machinery | 56.67 | Refusal Rate (%) | 97.1 |
Interactive version: theaggregate.ai/model?slug=gpt-3-5-turbo-0301 · How It Works · Data refreshed daily, snapshot 2026-11-08.