---
title: "Math leaderboard"
description: "Rankings on math derivation and calculation prompts."
canonical: https://pirouter.ai/rankings/math
last-updated: 2026-10-15
---

# Math leaderboard

> Rankings on math derivation and calculation prompts.

- Updated: 2026-09-30
- 966,903 votes · 371 models
- Score · higher is better

## Full ranking

| Rank | Rank spread | Model | Vendor | Score | Votes |
|---|---|---|---|---|---|
| 1 | 1–34 | Gemini 4 Argon (high) | Google | 1528 ±35 | 256 |
| 2 | 1–19 | Claude Opus 5 (high) | Anthropic | 1523 ±12 | 2703 |
| 3 | 1–21 | Claude Fable 5 (high) | Anthropic | 1522 ±14 | 1866 |
| 4 | 1–35 | [Claude Fable 5.1 \(max\)](https://pirouter.ai/models/anthropic/claude-fable-5.1) | Anthropic | 1518 ±25 | 523 |
| 5 | 1–22 | Claude Opus 4.6 (high) | Anthropic | 1517 ±10 | 3953 |
| 6 | 1–29 | [Gemini 3.8 Flash \(high\)](https://pirouter.ai/models/google/gemini-3.8-flash) | Google | 1516 ±17 | 1196 |
| 7 | 1–39 | Muse Spark 1.3 (max) | Meta | 1512 ±24 | 557 |
| 8 | 1–58 | [Claude Opus 5.5 \(high\)](https://pirouter.ai/models/anthropic/claude-opus-5.5) | Anthropic | 1510 ±39 | 202 |
| 9 | 1–38 | [Gemini 3.7 Flash \(high\)](https://pirouter.ai/models/google/gemini-3.7-flash) | Google | 1507 ±18 | 1054 |
| 10 | 1–34 | Claude Opus 4.7 (high) | Anthropic | 1504 ±11 | 3240 |
| 11 | 1–56 | [DeepSeek V4.1 Flash \(max\)](https://pirouter.ai/models/deepseek/deepseek-v4.1-flash) | DeepSeek | 1502 ±28 | 408 |
| 12 | 1–38 | [Gemini 3.6 Flash \(high\)](https://pirouter.ai/models/google/gemini-3.6-flash) | Google | 1502 ±14 | 1708 |
| 13 | 1–48 | [GLM 5.3 Flash](https://pirouter.ai/models/z-ai/glm-5.3-flash) | Z.ai | 1500 ±18 | 1069 |
| 14 | 1–48 | Kimi K3 (max) | Moonshot | 1500 ±17 | 1190 |
| 15 | 2–38 | [GPT-5.5](https://pirouter.ai/models/openai/gpt-5.5) | OpenAI | 1499 ±11 | 3493 |
| 16 | 1–49 | Qwen3.8 (max) | Alibaba | 1497 ±18 | 1111 |
| 17 | 2–47 | [Gemini 3.5 Flash \(high\)](https://pirouter.ai/models/google/gemini-3.5-flash) | Google | 1496 ±13 | 2331 |
| 18 | 4–48 | Claude Opus 4.8 (high) | Anthropic | 1494 ±11 | 2918 |
| 19 | 3–51 | [GPT-5.6 Sol \(xhigh\)](https://pirouter.ai/models/openai/gpt-5.6-sol) | OpenAI | 1493 ±14 | 1697 |
| 20 | 4–53 | Muse Spark 1.1 | Meta | 1491 ±15 | 1618 |
| 21 | 4–48 | [GPT-5.4 \(high\)](https://pirouter.ai/models/openai/gpt-5.4) | OpenAI | 1491 ±11 | 3379 |
| 22 | 1–58 | GLM 5.3 (max) | Z.ai | 1491 ±22 | 698 |
| 23 | 1–72 | [GPT-6 Astra \(max\)](https://pirouter.ai/models/openai/gpt-6-astra) | OpenAI | 1490 ±28 | 436 |
| 24 | 5–49 | Gemini 3.1 Pro Preview | Google | 1488 ±8 | 6277 |
| 25 | 4–56 | [GLM 5.2 \(max\)](https://pirouter.ai/models/z-ai/glm-5.2) | Z.ai | 1487 ±13 | 2014 |
| 26 | 1–95 | MiMo V2.6 Pro | Xiaomi | 1486 ±40 | 213 |
| 27 | 1–95 | Qwen3.7 Preview (max) | Alibaba | 1484 ±38 | 231 |
| 28 | 8–58 | MiMo V2.5 Pro | Xiaomi | 1480 ±11 | 3320 |
| 29 | 4–80 | Hy3 | Tencent | 1480 ±25 | 531 |
| 30 | 7–62 | Kimi K2.6 | Moonshot | 1479 ±13 | 2074 |
| 31 | 5–67 | [GPT-5.6 Terra \(xhigh\)](https://pirouter.ai/models/openai/gpt-5.6-terra) | OpenAI | 1479 ±15 | 1630 |
| 32 | 5–69 | Inkling | Thinky | 1479 ±15 | 1559 |
| 33 | 8–68 | Ernie 5.1 | Baidu | 1477 ±13 | 1982 |
| 34 | 11–65 | Gemini 3 Pro | Google | 1477 ±11 | 2759 |
| 35 | 4–93 | Qwen3.6 Preview (max) | Alibaba | 1476 ±29 | 380 |
| 36 | 12–69 | GLM 5.1 | Z.ai | 1475 ±12 | 2829 |
| 37 | 8–74 | [GPT-5.6 Luna \(xhigh\)](https://pirouter.ai/models/openai/gpt-5.6-luna) | OpenAI | 1475 ±14 | 1745 |
| 38 | 12–72 | Gemini 3 Flash | Google | 1474 ±13 | 2058 |
| 39 | 12–74 | [Claude Sonnet 5 \(high\)](https://pirouter.ai/models/anthropic/claude-sonnet-5) | Anthropic | 1474 ±13 | 2018 |
| 40 | 12–75 | Grok 4.5 | xAI | 1473 ±14 | 1727 |
| 41 | 12–74 | Claude Opus 4.5 20251101 High 32k | Anthropic | 1472 ±12 | 2335 |
| 42 | 17–72 | Kimi K2.5 (thinking) | Moonshot | 1471 ±9 | 4151 |
| 43 | 5–100 | Gemma 4 31b | Google | 1469 ±26 | 432 |
| 44 | 4–106 | MiMo V2.6 Flash | Xiaomi | 1469 ±32 | 301 |
| 45 | 12–83 | Qwen3.5 Preview (max) | Alibaba | 1468 ±16 | 1423 |
| 46 | 5–101 | Gemma 4 26b A4b | Google | 1468 ±28 | 389 |
| 47 | 20–80 | [DeepSeek V4 Pro 0813 \(high\)](https://pirouter.ai/models/deepseek/deepseek-v4-pro) | DeepSeek | 1466 ±12 | 2706 |
| 48 | 12–94 | Muse Spark | Meta | 1466 ±20 | 892 |
| 49 | 19–84 | Qwen3.7 Plus | Alibaba | 1466 ±14 | 1917 |
| 50 | 21–80 | Grok 4.20 Beta 0309 Reasoning | xAI | 1465 ±10 | 3529 |
| 51 | 23–82 | Claude Sonnet 4.6 | Anthropic | 1463 ±10 | 3775 |
| 52 | 23–81 | Claude Opus 4.5 | Anthropic | 1463 ±9 | 4535 |
| 53 | 21–97 | GPT-5.5 Instant | OpenAI | 1461 ±16 | 1528 |
| 54 | 26–95 | GPT-5.1 (high) | OpenAI | 1458 ±12 | 2577 |
| 55 | 20–104 | Qwen3.8 27b | Alibaba | 1458 ±20 | 849 |
| 56 | 26–97 | GPT-5.2 (high) | OpenAI | 1456 ±11 | 3111 |
| 57 | 27–97 | Grok 4.20 Multi Agent Beta 0309 | xAI | 1455 ±11 | 3381 |
| 58 | 32–95 | Claude Sonnet 4.5 20250929 High 32k | Anthropic | 1455 ±8 | 5044 |
| 59 | 26–101 | GPT-5.2 Chat Latest | OpenAI | 1454 ±13 | 2161 |
| 60 | 13–129 | [GPT-6 Sol \(max\)](https://pirouter.ai/models/openai/gpt-6-sol) | OpenAI | 1454 ±29 | 406 |
| 61 | 27–100 | Qwen3.6 Plus | Alibaba | 1454 ±12 | 2690 |
| 62 | 32–97 | Gemini 3 Flash | Google | 1454 ±9 | 4976 |
| 63 | 35–100 | Qwen3.5 397b A17b | Alibaba | 1452 ±9 | 4808 |
| 64 | 28–109 | MiMo V2 Pro | Xiaomi | 1451 ±15 | 1671 |
| 65 | 38–101 | Dola Seed 2.0 Pro | ByteDance | 1450 ±10 | 4332 |
| 66 | 30–112 | Grok 4.20 Beta1 | xAI | 1450 ±14 | 1674 |
| 67 | 12–149 | Muse Glimmer | Meta | 1448 ±39 | 223 |
| 68 | 39–104 | O3 2025 04 16 | OpenAI | 1448 ±10 | 3677 |
| 69 | 21–134 | [GPT-6 Luna \(max\)](https://pirouter.ai/models/openai/gpt-6-luna) | OpenAI | 1448 ±27 | 440 |
| 70 | 29–122 | Inkling Small | Thinky | 1447 ±18 | 1133 |
| 71 | 32–126 | Grok 4.6 (high) | xAI | 1446 ±18 | 1096 |
| 72 | 19–147 | Grok 4.7 (xhigh) | xAI | 1445 ±34 | 288 |
| 73 | 42–115 | Claude Opus 4.1 (thinking) | Anthropic | 1443 ±11 | 3045 |
| 74 | 39–121 | GLM 5 | Z.ai | 1443 ±14 | 1686 |
| 75 | 26–136 | Kimi K2.5 Instant | Moonshot | 1443 ±24 | 552 |
| 76 | 46–115 | Grok 4.1 (thinking) | xAI | 1442 ±10 | 3939 |
| 77 | 27–138 | Nvidia Nemotron 3 Ultra 550b A55b Nvfp4 | NVIDIA | 1442 ±24 | 576 |
| 78 | 39–127 | [Gemini 3.5 Flash Lite](https://pirouter.ai/models/google/gemini-3.5-flash-lite) | Google | 1442 ±15 | 1572 |
| 79 | 44–121 | [DeepSeek V4 Flash 0731 \(high\)](https://pirouter.ai/models/deepseek/deepseek-v4-flash) | DeepSeek | 1441 ±12 | 2618 |
| 80 | 48–115 | Gemini 2.5 Pro | Google | 1439 ±7 | 7766 |
| 81 | 46–125 | [GPT-5.4 Mini \(high\)](https://pirouter.ai/models/openai/gpt-5.4-mini) | OpenAI | 1439 ±11 | 3364 |
| 82 | 46–129 | MiMo V2.5 | Xiaomi | 1438 ±12 | 2454 |
| 83 | 52–127 | Kimi K2 Thinking Turbo | Moonshot | 1436 ±10 | 3905 |
| 84 | 35–147 | GLM 5v Turbo | Z.ai | 1436 ±25 | 482 |
| 85 | 52–129 | Gemini 3.1 Flash Lite Preview | Google | 1435 ±10 | 3617 |
| 86 | 46–134 | Longcat Flash Chat 2602 Exp | Meituan | 1435 ±14 | 1826 |
| 87 | 46–132 | Ernie 5.0 0110 | Baidu | 1435 ±12 | 2248 |
| 88 | 43–140 | MiMo V2 Omni | Xiaomi | 1434 ±19 | 977 |
| 89 | 46–136 | Qwen3 Preview (max) | Alibaba | 1434 ±15 | 1488 |
| 90 | 46–135 | GPT-5 (high) | OpenAI | 1434 ±14 | 1859 |
| 91 | 56–133 | Minimax M3 | MiniMax | 1433 ±12 | 2854 |
| 92 | 39–154 | Hunyuan Hy3 Preview | Tencent | 1431 ±27 | 429 |
| 93 | 56–139 | GPT-5.3 Chat Latest | OpenAI | 1431 ±13 | 2135 |
| 94 | 62–139 | DeepSeek V3.2 (thinking) | DeepSeek | 1428 ±12 | 2623 |
| 95 | 45–153 | Mistral Medium 3.5 | Mistral | 1428 ±24 | 565 |
| 96 | 46–152 | GLM 4.7 | Z.ai | 1428 ±20 | 734 |
| 97 | 66–136 | Claude Sonnet 4.5 | Anthropic | 1427 ±8 | 5103 |
| 98 | 28–169 | Amazon Nova Experimental Chat 26 02 10 | Amazon | 1427 ±39 | 211 |
| 99 | 62–141 | Qwen3.5 122b A10b | Alibaba | 1426 ±14 | 1836 |
| 100 | 62–140 | Grok 4 0709 | xAI | 1426 ±12 | 2233 |

## FAQ

### What is the best AI model for math?

As of 2026-09-30, Google Gemini 4 Argon (high) is #1 on the math board with a score of 1528, rank spread 1–34.

### What is the best open-weights model for math?

DeepSeek V4.1 Flash (DeepSeek), #11 on this board.

### What does this board rank?

Rankings on math derivation and calculation prompts.

### How often is it updated?

Whenever new models ship or the rankings change; this board was last updated on 2026-09-30.

Data: [LMArena leaderboard dataset](https://huggingface.co/datasets/lmarena-ai/leaderboard-dataset), used under [CC BY 4.0](https://creativecommons.org/licenses/by/4.0/); filtered and reasoning tiers merged, scores unchanged.
View the original board: https://arena.ai/leaderboard/text/math