BenchMark'd
No ratings yet
Be the first to review this model
A fast, cost-effective reasoning model from Amazon. Scores 79.6% on GPQA Diamond and 91% on AIME 2025. Designed for high-throughput applications that need capable reasoning at low cost.
Released
December 2, 2025
Parameters
Unknown
Context
1M
Pricing
Paid
| Benchmark | Category | Score | Performance |
|---|---|---|---|
MMLU | knowledge | 82.4% | 82 |
HumanEval | coding | 78.2% | 78 |
MATH | reasoning | 76.8% | 77 |
Chatbot Arena ELO | overall | 1310 | 73 |
Last updated: March 15, 2026
Benchmark scores may vary based on evaluation methodology and conditions.