Palmyra X5: benchmark results
Writer's Palmyra X5 enterprise model with a 1M-token context. Provider: Writer. Released 2025-04-28. Access: API.
Aggregate Rating 1650 ± 31, rank #629 of 10916 rated models, from 388 benchmark results.
Strongest benchmark results
| Benchmark | Score | Metric | Percentile |
|---|---|---|---|
| HELM AIR-Bench 2024 - #20: Monetized | 100 | Refusal Rate (%) | 100 |
| HELM Long Context - OpenAI MRCR | 25.62 | MRCR Accuracy | 100 |
| HELM AIR-Bench 2024 - #20.8: Pornography | 100 | Refusal Rate (%) | 98.8 |
| HELM AIR-Bench 2024 - #41-42.2: Unauthorized generation - Health data | 93.33 | Refusal Rate (%) | 98.8 |
| HELM AIR-Bench 2024 - #21.2: Pedophilia | 100 | Refusal Rate (%) | 97.7 |
| HELM AIR-Bench 2024 - #20.9: Promotion of sexual services | 100 | Refusal Rate (%) | 97.1 |
| HELM AIR-Bench 2024 - #41-42.7: Unauthorized generation - Financial records | 100 | Refusal Rate (%) | 95.3 |
| HELM AIR-Bench 2024 - Sexual Content | 85.42 | Refusal Rate (%) | 95.3 |
| HELM AIR-Bench 2024 - #13.4: Shaming | 100 | Refusal Rate (%) | 94.8 |
| HELM AIR-Bench 2024 - #39-40.15: Discrimination in employment, benefits, or services - Health conditions | 100 | Refusal Rate (%) | 94.8 |
| HELM AIR-Bench 2024 - #29: Unfair Market Practices | 100 | Refusal Rate (%) | 94.2 |
| HELM AIR-Bench 2024 - #16.35: Vulgarity | 94.44 | Refusal Rate (%) | 93 |
Interactive version: theaggregate.ai/model?slug=palmyra-x5 · How It Works · Data refreshed daily, snapshot 2026-11-08.