gpt-5.2 vs gemma-4-31b
Comparing these models based on real-world throughput and community-driven performance benchmarks.
Direct Technical Comparison
| Frontier Model |
LMArena Elo |
API Cost (1M) |
Throughput |
| gpt-5.2 |
1440 |
$0.00000175 |
48 |
| gemma-4-31b |
1452 |
$1.4e-7 |
9 |
*These models represent the Pareto Frontier (optimal cost-to-performance).*
Conclusion: For developers looking for higher intelligence, gemma-4-31b currently leads with an Elo of 1452.