LMArena Text: mathematical work leaderboard
- Gemini 4 Argon leads
- 1557
- models measured
- 152
- lead over #2
- +14 Elo
- median
- 1433
Leader on LMArena Text: mathematical work · read October 22
On October 22, 2026, 164 of our catalog's models had a LMArena Text: mathematical work result, and Google Gemini 4 Argon led with 1557, ahead of Anthropic Claude Opus 5 at 1544.
Prompts from mathematical occupations. People ask a question, compare two anonymous answers and vote for the better one; the rating is built from those votes, as LMArena publishes it.
Role in the index
LMArena Text: mathematical work is shown as published and does not feed any of our indexes. Read the methodology
| # | Model | Result | Votes | Interval (±) | Evidence | Coding index | Output $/Mtok |
|---|---|---|---|---|---|---|---|
| 1 | Gemini 4 ArgonGoogle | 1557 | 331 | ±31.0 | 81.0 | $10.00 | |
| 2 | Claude Opus 5Anthropic | 1544 | 3,925 | ±10.5 | 83.0 | $25.00 | |
| 3 | Claude Fable 5.1Anthropic | 1541 | 878 | ±20.1 | 84.2 | $50.00 | |
| 4 | Claude Opus 4.6Anthropic | 1525 | 4,308 | ±9.8 | 57.2 | $25.00 | |
| 5 | Gemini 3.8 FlashGoogle | 1519 | 2,495 | ±12.5 | 61.6 | $3.75 | |
| 6 | Claude Opus 5.5Anthropic | 1516 | 549 | ±24.2 | 88.8 | $20.00 | |
| 7 | Claude Fable 5Anthropic | 1515 | 2,528 | ±12.3 | 79.3 | $50.00 | |
| 8 | Muse WonderMeta | 1514 | 301 | ±32.8 | 71.9 | $20.00 | |
| 9 | GPT-6.1 SolOpenAI | 1513 | 405 | ±28.1 | 82.4 | $10.00 | |
| 10 | MiMo-V2.6-ProXiaomi | 1511 | 228 | ±38.4 | 65.4 | $0.87 | |
| 11 | Claude Sonnet 5.5Anthropic | 1510 | 445 | ±27.5 | 87.3 | $10.00 | |
| 12 | Claude Opus 4.7Anthropic | 1509 | 3,656 | ±10.6 | 67.5 | $25.00 | |
| 13 | Kimi K3Moonshot AI | 1505 | 1,811 | ±14.3 | 62.3 | $15.00 | |
| 14 | Gemini 3.7 FlashGoogle | 1502 | 1,791 | ±14.4 | 58.0 | $3.75 | |
| 15 | Claude Opus 4.8Anthropic | 1497 | 3,478 | ±10.6 | 65.0 | $25.00 | |
| 16 | GPT-5.4OpenAI | 1497 | 3,546 | ±10.6 | 64.4 | $15.00 | |
| 17 | GLM 5.3Z.ai | 1497 | 1,272 | ±16.6 | 70.6 | $4.40 | |
| 18 | GLM 5.3 FlashZ.ai | 1497 | 2,086 | ±13.3 | 57.7 | $0.50 | |
| 19 | GPT-5.6 SolOpenAI | 1496 | 2,334 | ±12.8 | 73.5 | $20.00 | |
| 20 | GPT-5.5OpenAI | 1496 | 3,975 | ±10.1 | 71.3 | $30.00 | |
| 21 | Gemini 3.5 FlashGoogle | 1495 | 2,695 | ±12.0 | 39.8 | $9.00 | |
| 22 | Gemini 3.6 FlashGoogle | 1492 | 2,135 | ±13.2 | 62.5 | $3.75 | |
| 23 | Muse Spark 1.3Meta | 1491 | 1,137 | ±17.5 | 57.4 | $4.25 | |
| 24 | Qwen3.8 Max (0902)Qwen | 1491 | 1,730 | ±14.4 | 62.4 | $6.00 | |
| 25 | MiMo-V2.5-ProXiaomi | 1491 | 4,199 | ±9.8 | 31.6 | $0.9152 | |
| 26 | GLM 5.2Z.ai | 1489 | 2,786 | ±11.6 | 61.1 | $4.40 | |
| 27 | Muse Spark 1.1Meta | 1487 | 2,577 | ±12.1 | 70.3 | $4.25 | |
| 28 | DeepSeek V4.1 FlashDeepSeek | 1485 | 865 | ±19.6 | 62.5 | $0.84 | |
| 29 | GPT-6 AstraOpenAI | 1484 | 775 | ±21.3 | 85.1 | $50.00 | |
| 30 | Claude Sonnet 4.6Anthropic | 1483 | 3,986 | ±10.1 | 57.3 | $15.00 | |
| 31 | Hy3Tencent | 1482 | 754 | ±21.1 | Not benchmarked | $0.528 | |
| 32 | Grok 4.5xAI | 1482 | 2,162 | ±13.2 | 58.6 | $6.00 | |
| 33 | Claude Sonnet 5Anthropic | 1481 | 2,774 | ±11.6 | 57.7 | $10.00 | |
| 34 | Muse Spark 1.2Meta | 1480 | 520 | ±25.9 | 55.7 | $4.25 | |
| 35 | Kimi K2.6Moonshot AI | 1480 | 2,195 | ±12.9 | 52.1 | $4.00 | |
| 36 | MiMo-V2.6-FlashXiaomi | 1478 | 756 | ±21.1 | 54.7 | $0.28 | |
| 37 | GPT-5.6 TerraOpenAI | 1478 | 2,598 | ±12.1 | 63.9 | $12.00 | |
| 38 | Gemini 3.1 Pro PreviewGoogle | 1478 | 7,274 | ±7.9 | 32.3 | $12.00 | |
| 39 | Gemma 4 26B A4BGoogle | 1475 | 311 | ±31.1 | Not benchmarked | $0.60 | |
| 40 | R Beam 1.0 501B A23BReflection | 1475 | 251 | ±36.1 | Not benchmarked | $1.10 | |
| 41 | Kimi K2.5Moonshot AI | 1473 | 4,063 | ±9.6 | 38.6 | $2.85 | |
| 42 | Qwen3.7 PlusQwen | 1470 | 2,251 | ±12.8 | 33.2 | $1.28 | |
| 43 | GLM 5.1Z.ai | 1470 | 3,505 | ±10.6 | 49.2 | $4.40 | |
| 44 | GPT-5.6 LunaOpenAI | 1470 | 2,318 | ±12.8 | 58.0 | $1.20 | |
| 45 | Claude Opus 4.5Anthropic | 1468 | 3,776 | ±9.9 | 44.0 | $25.00 | |
| 46 | Gemma 4 31BGoogle | 1465 | 345 | ±29.2 | 39.0 | $0.40 | |
| 47 | Grok 4.6xAI | 1464 | 1,840 | ±14.2 | 65.7 | $6.00 | |
| 48 | Claude Sonnet 4.5Anthropic | 1462 | 4,204 | ±9.3 | 39.5 | $15.00 | |
| 49 | Mistral Large 4Mistral AI | 1462 | 521 | ±25.0 | 56.4 | $4.18 | |
| 50 | Nemotron 3 UltraNVIDIA | 1461 | 715 | ±22.0 | 30.0 | $2.40 | |
| 51 | T InklingThinking Machines | 1460 | 2,352 | ±12.7 | 31.5 | $4.05 | |
| 52 | Step 5 PreviewStepFun | 1460 | 502 | ±25.4 | 61.1 | $2.70 | |
| 53 | Qwen3.8 27BQwen | 1458 | 1,462 | ±15.4 | 74.2 | $2.55 | |
| 54 | DeepSeek V4 Pro 0423DeepSeek | 1455 | 3,288 | ±10.9 | 38.7 | $2.976 | |
| 55 | Gemini 2.5 ProGoogle | 1454 | 6,914 | ±7.3 | 19.1 | $10.00 | |
| 56 | Qwen3.5 397B A17BQwen | 1454 | 5,556 | ±8.6 | Not benchmarked | $2.34 | |
| 57 | GLM 5Z.ai | 1451 | 1,593 | ±14.8 | 41.7 | $3.20 | |
| 58 | MiMo-V2.5Xiaomi | 1451 | 2,633 | ±12.2 | 50.0 | $0.336 | |
| 59 | GPT-5.1OpenAI | 1449 | 1,869 | ±13.4 | 40.4 | $10.00 | |
| 60 | MiniMax M3MiniMax | 1448 | 3,781 | ±10.3 | 39.2 | $1.20 | |
| 61 | Qwen3.6 PlusQwen | 1448 | 2,874 | ±11.5 | 44.1 | $1.95 | |
| 62 | GLM 5V TurboZ.ai | 1447 | 601 | ±23.4 | Not benchmarked | $4.00 | |
| 63 | GLM 4.6Z.ai | 1447 | 1,784 | ±13.7 | 8.8 | $2.20 | |
| 64 | DeepSeek V4 Pro 0813DeepSeek | 1443 | 513 | ±26.1 | 56.1 | $1.98 | |
| 65 | GPT-5.2OpenAI | 1442 | 2,624 | ±11.9 | 55.9 | $14.00 | |
| 66 | Hy3 previewTencent | 1442 | 452 | ±27.3 | Not benchmarked | $0.60 | |
| 67 | T Inkling SmallThinking Machines | 1441 | 1,857 | ±14.1 | 27.3 | $1.20 | |
| 68 | Mistral Medium 3.5Mistral AI | 1440 | 648 | ±22.9 | 25.5 | $7.50 | |
| 69 | DeepSeek V3.2DeepSeek | 1440 | 1,968 | ±13.3 | 30.0 | $0.88 | |
| 70 | MiniMax M2.7MiniMax | 1437 | 4,620 | ±9.4 | 36.2 | $1.20 | |
| 71 | Qwen3 235B A22B Thinking 2507Qwen | 1435 | 476 | ±25.7 | Not benchmarked | $3.50 | |
| 72 | Grok 4.7xAI | 1435 | 773 | ±20.6 | 66.7 | $6.00 | |
| 73 | Qwen3 Next 80B A3B InstructQwen | 1434 | 1,078 | ±17.6 | Not benchmarked | $1.10 | |
| 74 | Claude Opus 4.1Anthropic | 1434 | 2,485 | ±11.6 | Not benchmarked | $75.00 | |
| 75 | Gemini 3.5 Flash LiteGoogle | 1433 | 2,033 | ±13.5 | 49.0 | $2.50 | |
| 76 | DeepSeek V3.2 ExpDeepSeek | 1433 | 404 | ±29.1 | Not benchmarked | $0.41 | |
| 77 | DeepSeek V4 Flash 0423DeepSeek | 1432 | 2,972 | ±11.4 | Not benchmarked | $0.28 | |
| 78 | Qwen3.5-122B-A10BQwen | 1431 | 1,721 | ±14.4 | Not benchmarked | $2.08 | |
| 79 | GLM 4.5Z.ai | 1430 | 1,260 | ±16.9 | Not benchmarked | $2.20 | |
| 80 | Claude Haiku 4.5Anthropic | 1430 | 8,088 | ±7.2 | 35.6 | $5.00 | |
| 81 | Qwen3.5-27BQwen | 1429 | 1,670 | ±14.7 | Not benchmarked | $1.56 | |
| 82 | GPT-5.4 MiniOpenAI | 1429 | 3,431 | ±10.8 | 35.8 | $4.50 | |
| 83 | Solar Pro 4Upstage | 1427 | 1,360 | ±16.3 | Not benchmarked | $0.60 | |
| 84 | GPT-6 LunaOpenAI | 1427 | 3,897 | ±10.6 | 57.3 | $0.50 | |
| 85 | GPT-6 SolOpenAI | 1424 | 873 | ±20.0 | 76.0 | $10.00 | |
| 86 | o3OpenAI | 1423 | 3,456 | ±10.3 | Not benchmarked | $8.00 | |
| 87 | Gemini 2.5 FlashGoogle | 1422 | 7,071 | ±7.2 | Not benchmarked | $2.50 | |
| 88 | Step 3.5 FlashStepFun | 1421 | 3,157 | ±10.8 | Not benchmarked | $0.30 | |
| 89 | Qwen3 32BQwen | 1420 | 282 | ±31.6 | Not benchmarked | $0.40 | |
| 90 | Qwen3 VL 235B A22B InstructQwen | 1419 | 600 | ±23.8 | Not benchmarked | $1.90 | |
| 91 | Gemini 3.1 Flash Lite PreviewGoogle | 1417 | 3,581 | ±10.4 | 22.7 | $1.50 | |
| 92 | R1 0528DeepSeek | 1415 | 919 | ±19.3 | Not benchmarked | $2.18 | |
| 93 | Qwen3.5-FlashQwen | 1413 | 3,356 | ±10.8 | 41.1 | $0.26 | |
| 94 | Mistral Medium 3.1Mistral AI | 1412 | 5,101 | ±8.4 | Not benchmarked | $2.00 | |
| 95 | Qwen3 30B A3B Instruct 2507Qwen | 1411 | 1,208 | ±16.4 | Not benchmarked | $0.30 | |
| 96 | GLM 4.7Z.ai | 1411 | 624 | ±22.3 | 55.1 | $2.20 | |
| 97 | Qwen3.5-35B-A3BQwen | 1410 | 1,726 | ±14.3 | Not benchmarked | $1.30 | |
| 98 | GPT-5.4 NanoOpenAI | 1409 | 3,478 | ±10.8 | 44.8 | $1.25 | |
| 99 | Nemotron 3 SuperNVIDIA | 1405 | 446 | ±26.1 | Not benchmarked | $0.45 | |
| 100 | GPT-5OpenAI | 1405 | 1,767 | ±14.2 | 38.8 | $10.00 | |
| 101 | o3 MiniOpenAI | 1399 | 1,711 | ±13.9 | Not benchmarked | $4.40 | |
| 102 | Grok 4.3xAI | 1396 | 3,936 | ±10.2 | 37.4 | $2.50 | |
| 103 | GLM 4.5 AirZ.ai | 1396 | 1,464 | ±15.2 | Not benchmarked | $1.10 | |
| 104 | Qwen3 Next 80B A3B ThinkingQwen | 1390 | 687 | ±22.2 | Not benchmarked | $1.20 | |
| 105 | o1OpenAI | 1390 | 2,644 | ±11.5 | Not benchmarked | $60.00 | |
| 106 | R1RetiredDeepSeek | 1389 | 1,504 | ±14.7 | Not benchmarked | $2.50 | |
| 107 | gpt-oss-120bOpenAI | 1386 | 1,550 | ±14.7 | Not benchmarked | $0.50 | |
| 108 | GPT-5 MiniOpenAI | 1380 | 1,304 | ±16.4 | 34.2 | $2.00 | |
| 109 | o4 MiniOpenAI | 1379 | 2,781 | ±10.9 | Not benchmarked | $4.40 | |
| 110 | MiniMax M2.5MiniMax | 1378 | 2,408 | ±12.4 | 37.2 | $1.20 | |
| 111 | Claude Opus 4RetiredAnthropic | 1376 | 2,621 | ±11.5 | Not benchmarked | $75.00 | |
| 112 | Nemotron 3 Nano 30B A3BNVIDIA | 1375 | 730 | ±21.3 | Not benchmarked | $0.22 | |
| 113 | GLM 4.7 FlashZ.ai | 1369 | 515 | ±24.6 | Not benchmarked | $0.40 | |
| 114 | Nova 2 LiteAmazon | 1369 | 644 | ±23.2 | Not benchmarked | $2.50 | |
| 115 | MiniMax M2MiniMax | 1368 | 301 | ±32.9 | Not benchmarked | $1.02 | |
| 116 | Qwen3 Coder 480B A35BQwen | 1367 | 1,444 | ±15.5 | Not benchmarked | $1.50 | |
| 117 | o1-miniRetiredOpenAI | 1366 | 6,318 | ±8.4 | Not benchmarked | $4.40 | |
| 118 | Qwen3 30B A3BQwen | 1363 | 1,577 | ±14.4 | Not benchmarked | $0.50 | |
| 119 | MiniMax M1MiniMax | 1363 | 1,767 | ±13.8 | Not benchmarked | $2.20 | |
| 120 | GPT-4.1OpenAI | 1360 | 3,085 | ±10.5 | Not benchmarked | $8.00 | |
| 121 | Gemini 2.0 FlashRetiredGoogle | 1357 | 3,748 | ±9.4 | Not benchmarked | $0.40 | |
| 122 | Mistral Medium 3Mistral AI | 1354 | 2,193 | ±12.2 | Not benchmarked | $2.00 | |
| 123 | Trinity Large ThinkingArcee AI | 1350 | 1,755 | ±15.3 | Not benchmarked | $0.80 | |
| 124 | GPT-4.1 MiniOpenAI | 1347 | 2,477 | ±11.6 | Not benchmarked | $1.60 | |
| 125 | Gemma 3 12BGoogle | 1336 | 333 | ±30.0 | Not benchmarked | $0.1625 | |
| 126 | GLM 4.5VZ.ai | 1336 | 226 | ±37.5 | Not benchmarked | $1.80 | |
| 127 | Qwen-PlusQwen | 1332 | 627 | ±21.4 | Not benchmarked | $0.78 | |
| 128 | Gemma 3 27BGoogle | 1327 | 3,210 | ±10.1 | Not benchmarked | $0.30 | |
| 129 | GPT-5 NanoOpenAI | 1326 | 440 | ±28.4 | Not benchmarked | $0.40 | |
| 130 | Gemini 1.5 ProRetiredGoogle | 1324 | 6,506 | ±7.8 | Not benchmarked | $5.00 | |
| 131 | Claude 3.7 SonnetRetiredAnthropic | 1314 | 3,176 | ±10.3 | Not benchmarked | $15.00 | |
| 132 | Command ACohere | 1312 | 3,654 | ±9.6 | Not benchmarked | $10.00 | |
| 133 | Claude 3.5 SonnetRetiredAnthropic | 1307 | 8,596 | ±7.0 | Not benchmarked | $15.00 | |
| 134 | gpt-oss-20bOpenAI | 1305 | 585 | ±24.2 | Not benchmarked | $0.15 | |
| 135 | Llama 4 MaverickMeta Llama | 1295 | 2,569 | ±11.5 | Not benchmarked | $1.00 | |
| 136 | Qwen2.5 72B InstructQwen | 1290 | 4,565 | ±8.8 | Not benchmarked | $0.40 | |
| 137 | Llama 4 ScoutMeta Llama | 1286 | 1,804 | ±13.7 | Not benchmarked | $0.50 | |
| 138 | GPT-4o (2024-05-13)OpenAI | 1284 | 13,306 | ±7.3 | Not benchmarked | $15.00 | |
| 139 | GPT-4o (2024-08-06)OpenAI | 1283 | 5,805 | ±8.5 | Not benchmarked | $10.00 | |
| 140 | Gemini 1.5 FlashRetiredGoogle | 1279 | 4,114 | ±9.3 | Not benchmarked | $0.30 | |
| 141 | GPT-4.1 NanoOpenAI | 1274 | 534 | ±23.9 | Not benchmarked | $0.40 | |
| 142 | GPT-4o-mini (2024-07-18)OpenAI | 1271 | 7,966 | ±7.4 | Not benchmarked | $0.60 | |
| 143 | Mistral Small 3.1 24BMistral AI | 1271 | 1,933 | ±13.5 | Not benchmarked | $0.555 | |
| 144 | GPT-4 TurboOpenAI | 1268 | 11,796 | ±8.0 | Not benchmarked | $30.00 | |
| 145 | Llama 3.3 70B InstructMeta Llama | 1267 | 5,237 | ±8.1 | Not benchmarked | $0.72 | |
| 146 | Claude 3 OpusRetiredAnthropic | 1264 | 22,970 | ±6.8 | Not benchmarked | $75.00 | |
| 147 | Nova Pro 1.0Amazon | 1261 | 2,650 | ±10.5 | Not benchmarked | $3.20 | |
| 148 | Mistral LargeMistral AI | 1257 | 3,037 | ±10.1 | Not benchmarked | $6.00 | |
| 149 | Gemma 3 4BGoogle | 1256 | 365 | ±29.9 | Not benchmarked | $0.075 | |
| 150 | Phi 4Microsoft | 1254 | 2,350 | ±11.6 | Not benchmarked | $0.14 | |
| 151 | Llama 3.1 70B InstructMeta Llama | 1253 | 6,524 | ±8.1 | Not benchmarked | $0.56 | |
| 152 | Claude 3.5 HaikuRetiredAnthropic | 1249 | 5,756 | ±8.0 | Not benchmarked | $4.00 | |
| 153 | Qwen2.5 Coder 32B InstructQwen | 1243 | 619 | ±20.7 | Not benchmarked | $1.00 | |
| 154 | Mistral Small 3Mistral AI | 1243 | 1,447 | ±14.4 | Not benchmarked | $0.08 | |
| 155 | Nova Lite 1.0Amazon | 1235 | 2,173 | ±11.7 | Not benchmarked | $0.24 | |
| 156 | GPT-4OpenAI | 1228 | 6,112 | ±10.2 | Not benchmarked | $60.00 | |
| 157 | Gemma 2 27BGoogle | 1221 | 8,869 | ±7.2 | Not benchmarked | $0.65 | |
| 158 | Claude 3 SonnetRetiredAnthropic | 1213 | 12,413 | ±8.4 | Not benchmarked | $15.00 | |
| 159 | Nova Micro 1.0Amazon | 1212 | 2,016 | ±12.0 | Not benchmarked | $0.14 | |
| 160 | Command R (08-2024)Cohere | 1198 | 1,243 | ±14.9 | Not benchmarked | $0.60 | |
| 161 | Claude 3 HaikuRetiredAnthropic | 1187 | 13,491 | ±7.8 | Not benchmarked | $1.25 | |
| 162 | Llama 3.1 8B InstructMeta Llama | 1180 | 5,985 | ±8.5 | Not benchmarked | $0.08 | |
| 163 | Llama 3.2 3B InstructMeta Llama | 1128 | 980 | ±16.7 | Not benchmarked | $0.3325 | |
| 164 | Llama 3.2 1B InstructMeta Llama | 1098 | 976 | ±17.8 | Not benchmarked | $0.201 |
Only models in our catalog are listed, and each appears once with its best published run. 184 rows were read from the source on October 22, 2026, and rows that match no catalog model are left out. Every other column is a figure the board publishes for the same run.