GLM 5.2 is the most-used model for math on OpenRouter, handling 8.9% of requests classified as math over the last 7 days.
This ranks models by how often they are actually chosen for math in production traffic, not by benchmark score. Adoption reflects price, latency and availability as much as raw capability — which is exactly why it often disagrees with the benchmark leaderboards, and why it is worth reading alongside them.
| # | Model | Request share | Token share | Input /M | Context |
|---|---|---|---|---|---|
| 1 | 8.9% | 7.5% | $0.760 | 1.0M | |
| 2 | 7.9% | 11.8% | $0.140 | 1.0M | |
| 3 | 5.1% | 2.2% | $0.300 | 1.0M | |
| 4 | 3.9% | 1.5% | $0.048 | 262K | |
| 5 | 3.4% | 2.4% | $0.037 | 131K | |
| 6 | 3.0% | 1.8% | $0.030 | 131K | |
| 7 | 2.5% | 6.1% | $0.140 | 1.0M | |
| 8 | 2.3% | 1.6% | $0.500 | 1.0M | |
| 9 | 2.3% | 6.7% | $0.435 | 1.0M | |
| 10 | 2.2% | 0.40% | $0.050 | 131K |
Reading the two columns together: DeepSeek V4 Flash 0423, DeepSeek V4 Pro take a noticeably larger share of tokens than of requests — meaning they are being used for the longer, heavier math jobs rather than quick one-shot calls.
GLM 5.2 is the most-used model for math on OpenRouter, handling 8.9% of requests classified as math over the last 7 days. It is followed by DeepSeek V4 Flash 0423 (7.9%) and Gemini 2.5 Flash (5.1%).
No. This ranks by how often each model is actually chosen for math in production traffic through OpenRouter — real-world adoption, which reflects price and availability as much as capability. For capability-based rankings, see the benchmark leaderboards.
Math accounts for 0.30% of classified requests and 0.10% of classified tokens on OpenRouter over the trailing 7 days.
Source: OpenRouter (openrouter.ai/rankings), as of 2026-08-04. Shares are of classified, sampled traffic over a trailing 7-day window; absolute volumes are not published.