Рейтинг моделей для кода
Таблица по открытым замерам оценок людей: оценка, интервал и число голосов; переключайте категорию и дату снимка.
Динамика рейтинга
Загружаем историю замеров…
Как считается
Оценки и интервалы приведены как в открытых замерах оценок людей, без пересчёта; доля использования — по публичным данным API-агрегаторов (топ-K моделей и строка «прочие»); характеристики и цены — из каталогов и провайдеров с датой наблюдения. Интервал у оценки не равен диапазону мест: соседние модели с пересекающимися интервалами статистически неразличимы.
Полная методика: источники, частота обновления, что означает число
Смысл числа
- Метрика
- рейтинг Брэдли — Терри (Elo-подобная шкала) с интервалом из источника
- Единица выборки
- голос в парном сравнении
- Окно
- снимок датасета от 13 сентября 2026
- Как читать
- интервал не равен диапазону мест: модели с пересекающимися интервалами статистически неразличимы
Таблица борда
Снимок от 13 сентября 2026, категория «код», 300 строк. Оценка и интервал — из источника; интервал не равен диапазону мест.
| Место | Модель | Разработчик | Оценка | Интервал | Голоса | Лицензия |
|---|---|---|---|---|---|---|
| 1 | Claude Opus 4.6 в режиме high | Anthropic | 1 535 | 1 530–1 541 | 18 766 | Proprietary |
| 2 | Claude Opus 4.6 | Anthropic | 1 534 | 1 529–1 539 | 21 152 | Proprietary |
| 3 | Claude Opus 5 в режиме high | Anthropic | 1 532 | 1 525–1 539 | 11 403 | Proprietary |
| 4 | Claude Opus 5 в режиме max | Anthropic | 1 530 | 1 522–1 539 | 5 404 | Proprietary |
| 5 | Claude Fable 5 | Anthropic | 1 520 | 1 512–1 527 | 8 040 | Proprietary |
| 6 | Claude Opus 4.7 в режиме high | Anthropic | 1 519 | 1 513–1 525 | 17 149 | Proprietary |
| 7 | Muse Spark 1.3 в режиме max | Meta | 1 516 | 1 499–1 532 | 1 346 | Proprietary |
| 8 | Claude Opus 4.7 | Anthropic | 1 516 | 1 510–1 522 | 17 321 | Proprietary |
| 9 | Gemini 3.8 Flash в режиме high | 1 514 | 1 498–1 530 | 1 434 | Proprietary | |
| 10 | Claude Fable 5.1 в режиме max | Anthropic | 1 511 | 1 494–1 529 | 1 184 | Proprietary |
| 11 | GLM 5.3 Flash | Z.ai | 1 508 | 1 497–1 520 | 2 755 | MIT |
| 12 | Kimi K3 в режиме max | Moonshot AI | 1 507 | 1 499–1 516 | 5 507 | Kimi K3 license |
| 13 | Claude Sonnet 4.6 | Anthropic | 1 504 | 1 498–1 510 | 18 320 | Proprietary |
| 14 | Claude Opus 4.5 в режиме high | Anthropic | 1 503 | 1 496–1 510 | 7 515 | Proprietary |
| 15 | Gemini 3.7 Flash в режиме high | 1 503 | 1 488–1 517 | 1 730 | Proprietary | |
| 16 | MiMo-V2.5-Pro | Xiaomi | 1 501 | 1 495–1 507 | 16 683 | MIT |
| 17 | GLM 5.3 в режиме max | Z.ai | 1 500 | 1 489–1 512 | 2 846 | MIT |
| 18 | Qwen3.8 Max (0902) | Alibaba Qwen | 1 500 | 1 491–1 509 | 4 804 | Proprietary |
| 19 | Claude Opus 4.5 | Anthropic | 1 499 | 1 494–1 504 | 17 174 | Proprietary |
| 20 | Muse Spark 1.2 в режиме xhigh | Meta | 1 499 | 1 479–1 518 | 928 | Proprietary |
| 21 | Qwen3.7 Max | Alibaba Qwen | 1 498 | 1 479–1 516 | 1 117 | Proprietary |
| 22 | Muse Spark 1.1 | Meta | 1 496 | 1 488–1 503 | 7 958 | Proprietary |
| 23 | GPT-5.4 в режиме high | OpenAI | 1 496 | 1 490–1 501 | 16 395 | Proprietary |
| 24 | GPT-5.5 в режиме high | OpenAI | 1 495 | 1 489–1 501 | 18 031 | Proprietary |
| 25 | Gemini 3.5 Flash в режиме high | 1 491 | 1 484–1 497 | 11 109 | Proprietary | |
| 26 | Gemini 3.6 Flash в режиме high | 1 491 | 1 483–1 498 | 7 795 | Proprietary | |
| 27 | Claude Opus 4.8 в режиме high | Anthropic | 1 491 | 1 484–1 497 | 14 473 | Proprietary |
| 28 | GPT-5.6 Sol в режиме xhigh | OpenAI | 1 490 | 1 482–1 497 | 7 547 | Proprietary |
| 29 | ERNIE 5.1 | Baidu | 1 489 | 1 482–1 496 | 10 269 | Proprietary |
| 30 | Claude Sonnet 4.5 в режиме high | Anthropic | 1 489 | 1 484–1 494 | 19 113 | Proprietary |
| 31 | GPT-6 Astra в режиме max | OpenAI | 1 489 | 1 465–1 512 | 645 | Proprietary |
| 32 | Claude Opus 4.8 | Anthropic | 1 487 | 1 481–1 494 | 14 998 | Proprietary |
| 33 | Qwen3.5 Max | Alibaba Qwen | 1 487 | 1 479–1 495 | 6 054 | Proprietary |
| 34 | Kimi K2.6 | Moonshot AI | 1 486 | 1 479–1 493 | 10 398 | Modified MIT |
| 35 | Claude Sonnet 4.5 | Anthropic | 1 486 | 1 481–1 491 | 18 939 | Proprietary |
| 36 | GLM 5.1 | Z.ai | 1 486 | 1 480–1 492 | 13 597 | MIT |
| 37 | Amazon Nova Experimental Chat · 26-02-10 | Amazon | 1 485 | 1 466–1 505 | 836 | Proprietary |
| 38 | GPT-5.6 Terra в режиме xhigh | OpenAI | 1 485 | 1 478–1 493 | 7 783 | Proprietary |
| 39 | Kimi K2.5 · instant | Moonshot AI | 1 485 | 1 471–1 498 | 1 769 | Modified MIT |
| 40 | Claude Sonnet 5 в режиме high | Anthropic | 1 485 | 1 478–1 492 | 9 725 | Proprietary |
| 41 | Gemini 3.5 Flash в режиме medium | 1 484 | 1 477–1 490 | 10 711 | Proprietary | |
| 42 | Gemini 3 Pro | 1 482 | 1 475–1 489 | 8 476 | Proprietary | |
| 43 | Gemini 3.1 Pro Preview | 1 482 | 1 477–1 487 | 29 544 | Proprietary | |
| 44 | Qwen3.8 27B | Alibaba Qwen | 1 482 | 1 471–1 492 | 3 135 | Apache 2.0 |
| 45 | GLM 5.2 в режиме max | Z.ai | 1 481 | 1 474–1 488 | 10 311 | MIT |
| 46 | GPT-5.5 · 5.5 | OpenAI | 1 481 | 1 475–1 487 | 18 748 | Proprietary |
| 47 | GPT-5.4 | OpenAI | 1 480 | 1 474–1 486 | 17 669 | Proprietary |
| 48 | Claude Opus 4.1 · 20250805-thinking-16k | Anthropic | 1 480 | 1 473–1 486 | 9 718 | Proprietary |
| 49 | Muse Spark | Meta | 1 479 | 1 468–1 489 | 3 771 | Proprietary |
| 50 | Grok 4.5 | xAI | 1 478 | 1 471–1 486 | 8 472 | Proprietary |
| 51 | MiMo-V2 Pro | Xiaomi | 1 476 | 1 468–1 484 | 6 913 | Proprietary |
| 52 | Kimi K2.5 · thinking | Moonshot AI | 1 474 | 1 468–1 479 | 19 559 | Modified MIT |
| 53 | Claude Opus 4.1 · 20250805 | Anthropic | 1 473 | 1 468–1 479 | 15 379 | Proprietary |
| 54 | Qwen3.7 Plus | Alibaba Qwen | 1 473 | 1 467–1 480 | 11 416 | Proprietary |
| 55 | LongCat-Flash-Chat · chat-2602-exp | Meituan | 1 473 | 1 466–1 480 | 7 749 | Proprietary |
| 56 | Seed 2.0 Pro | ByteDance Seed | 1 472 | 1 466–1 477 | 20 697 | Proprietary |
| 57 | MiniMax M3 | MiniMax | 1 471 | 1 465–1 477 | 14 110 | MiniMax Community License |
| 58 | DeepSeek V4 Pro 0423 | DeepSeek | 1 470 | 1 464–1 476 | 16 143 | MIT |
| 59 | MiMo-V2.5 | Xiaomi | 1 469 | 1 462–1 475 | 12 829 | MIT |
| 60 | Nemotron 3 Ultra | NVIDIA | 1 468 | 1 457–1 480 | 2 944 | OpenMDW-1.1 |
| 61 | Qwen3.6 Max Preview | Alibaba Qwen | 1 468 | 1 452–1 483 | 1 547 | Proprietary |
| 62 | GLM 5V Turbo | Z.ai | 1 468 | 1 456–1 479 | 2 612 | Proprietary |
| 63 | LongCat-Flash-Chat · chat | Meituan | 1 467 | 1 455–1 480 | 2 202 | MIT |
| 64 | Hy3 | Tencent | 1 467 | 1 454–1 480 | 2 259 | Apache 2.0 |
| 65 | DeepSeek V4 Pro 0813 в режиме high | DeepSeek | 1 467 | 1 455–1 479 | 2 593 | MIT |
| 66 | Qwen3.6 Plus | Alibaba Qwen | 1 466 | 1 460–1 472 | 13 758 | Proprietary |
| 67 | MiMo-V2 Omni | Xiaomi | 1 465 | 1 456–1 474 | 5 520 | Proprietary |
| 68 | Qwen3.5 397B A17B | Alibaba Qwen | 1 464 | 1 459–1 469 | 22 136 | Apache 2.0 |
| 69 | Grok 4.6 в режиме high | xAI | 1 463 | 1 454–1 473 | 4 012 | Proprietary |
| 70 | Inkling | Thinking Machines Lab | 1 463 | 1 456–1 471 | 7 414 | Apache 2.0 |
| 71 | GPT-5.6 Luna в режиме xhigh | OpenAI | 1 463 | 1 455–1 470 | 7 986 | Proprietary |
| 72 | Gemini 3 Flash Preview | 1 462 | 1 454–1 469 | 6 328 | Proprietary | |
| 73 | Mistral Medium 3.5 | Mistral AI | 1 461 | 1 450–1 472 | 3 070 | Modified MIT |
| 74 | GLM 5 | Z.ai | 1 461 | 1 453–1 468 | 7 179 | MIT |
| 75 | Grok 4.20 · beta-0309-reasoning | xAI | 1 458 | 1 452–1 464 | 17 193 | Proprietary |
| 76 | Grok 4.20 Multi-Agent | xAI | 1 457 | 1 451–1 462 | 16 894 | Proprietary |
| 77 | Qwen3 Max · preview | Alibaba Qwen | 1 456 | 1 448–1 465 | 5 307 | Proprietary |
| 78 | GLM 4.7 | Z.ai | 1 456 | 1 444–1 468 | 2 389 | MIT |
| 79 | DeepSeek V4 Flash 0423 | DeepSeek | 1 456 | 1 449–1 462 | 14 310 | MIT |
| 80 | Gemini 3.5 Flash Lite | 1 456 | 1 448–1 463 | 7 325 | Proprietary | |
| 81 | Gemma 4 31B | 1 455 | 1 440–1 471 | 1 373 | Apache 2.0 | |
| 82 | ERNIE 5.0 · 0110 | Baidu | 1 454 | 1 448–1 461 | 8 553 | Proprietary |
| 83 | MiniMax M2.7 | MiniMax | 1 454 | 1 449–1 460 | 19 870 | Modified MIT |
| 84 | Claude Haiku 4.5 | Anthropic | 1 454 | 1 450–1 458 | 33 262 | Proprietary |
| 85 | DeepSeek V3.2 · v3.2-thinking | DeepSeek | 1 453 | 1 447–1 460 | 8 481 | MIT |
| 86 | Kimi K2 Thinking | Moonshot AI | 1 453 | 1 448–1 459 | 14 683 | Modified MIT |
| 87 | DeepSeek V4 Pro 0423 в режиме high | DeepSeek | 1 453 | 1 447–1 459 | 15 077 | MIT |
| 88 | GPT-5.1 в режиме high | OpenAI | 1 453 | 1 446–1 460 | 8 106 | Proprietary |
| 89 | Gemini 2.5 Pro | 1 452 | 1 448–1 457 | 26 265 | Proprietary | |
| 90 | Inkling Small | Thinking Machines Lab | 1 452 | 1 443–1 460 | 5 384 | Apache 2.0 |
| 91 | DeepSeek V3.2 · v3.2 | DeepSeek | 1 449 | 1 442–1 455 | 10 510 | MIT |
| 92 | GLM 4.6 | Z.ai | 1 448 | 1 441–1 456 | 7 391 | MIT |
| 93 | Amazon Nova Experimental Chat · 26-01-10 | Amazon | 1 448 | 1 427–1 469 | 734 | Proprietary |
| 94 | GPT-5.2 Chat | OpenAI | 1 446 | 1 439–1 453 | 9 089 | Proprietary |
| 95 | Mistral Large 3 2512 | Mistral AI | 1 446 | 1 441–1 451 | 17 344 | Apache 2.0 |
| 96 | Grok 4.20 · beta1 | xAI | 1 446 | 1 439–1 454 | 7 156 | Proprietary |
| 97 | Qwen3 235B A22B Instruct 2507 | Alibaba Qwen | 1 445 | 1 440–1 449 | 21 114 | Apache 2.0 |
| 98 | Gemma 4 26B A4B | 1 445 | 1 429–1 460 | 1 377 | Apache 2.0 | |
| 99 | Grok 4.1 · 4.1-thinking | xAI | 1 444 | 1 439–1 450 | 14 831 | Proprietary |
| 100 | Grok 4.1 · 4.1 | xAI | 1 444 | 1 439–1 450 | 15 425 | Proprietary |
| 101 | Claude Opus 4 · 20250514-thinking-16k | Anthropic | 1 442 | 1 434–1 450 | 6 559 | Proprietary |
| 102 | GPT-5.2 в режиме high | OpenAI | 1 442 | 1 436–1 448 | 11 674 | Proprietary |
| 103 | MiMo-V2 Flash · flash (non-thinking) | Xiaomi | 1 441 | 1 435–1 447 | 11 904 | MIT |
| 104 | Qwen3 Next 80B A3B Instruct | Alibaba Qwen | 1 441 | 1 432–1 449 | 4 735 | Apache 2.0 |
| 105 | Qwen3 VL 235B A22B Instruct | Alibaba Qwen | 1 440 | 1 427–1 453 | 2 287 | Apache 2.0 |
| 106 | DeepSeek V4 Flash 0423 в режиме high | DeepSeek | 1 440 | 1 433–1 446 | 14 280 | MIT |
| 107 | DeepSeek V3.2 Exp · exp-thinking | DeepSeek | 1 439 | 1 426–1 453 | 1 906 | MIT |
| 108 | GPT-5.4 Mini в режиме high | OpenAI | 1 439 | 1 433–1 445 | 16 192 | Proprietary |
| 109 | Qwen3 Max · 2025-09-23 | Alibaba Qwen | 1 439 | 1 426–1 452 | 2 007 | Proprietary |
| 110 | GPT-5.1 | OpenAI | 1 437 | 1 430–1 443 | 9 023 | Proprietary |
| 111 | Gemini 3 Flash Preview в режиме minimal | 1 437 | 1 431–1 442 | 22 520 | Proprietary | |
| 112 | Step 3.5 Flash | StepFun | 1 437 | 1 431–1 442 | 14 294 | Apache 2.0 |
| 113 | Qwen3.5-122B-A10B | Alibaba Qwen | 1 436 | 1 429–1 444 | 7 777 | Apache 2.0 |
| 114 | GPT-5 в режиме high | OpenAI | 1 436 | 1 428–1 444 | 6 306 | Proprietary |
| 115 | GPT-5.5 · 5.5-instant | OpenAI | 1 434 | 1 426–1 442 | 7 507 | Proprietary |
| 116 | Mistral Medium 3.1 | Mistral AI | 1 434 | 1 429–1 439 | 20 916 | Proprietary |
| 117 | GLM 4.5 | Z.ai | 1 434 | 1 425–1 443 | 4 713 | MIT |
| 118 | DeepSeek V3.2 Exp · exp | DeepSeek | 1 433 | 1 421–1 445 | 2 471 | MIT |
| 119 | Grok 3 | xAI | 1 432 | 1 424–1 441 | 5 350 | Proprietary |
| 120 | GPT-5.2 | OpenAI | 1 432 | 1 427–1 437 | 20 773 | Proprietary |
| 121 | Amazon Nova Experimental Chat · 11-10 | Amazon | 1 432 | 1 424–1 440 | 5 267 | Proprietary |
| 122 | Grok 4 Fast · chat | xAI | 1 430 | 1 413–1 446 | 1 179 | Proprietary |
| 123 | Hy3 preview | Tencent | 1 429 | 1 415–1 443 | 1 963 | tencent-hunyuan-community |
| 124 | Amazon Nova Experimental Chat · 12-10 | Amazon | 1 428 | 1 407–1 449 | 707 | Proprietary |
| 125 | ERNIE 5.0 · preview-1203 | Baidu | 1 428 | 1 414–1 441 | 1 929 | Proprietary |
| 126 | DeepSeek V3.1 Terminus · terminus-thinking | DeepSeek | 1 427 | 1 404–1 450 | 632 | MIT |
| 127 | R1 0528 | DeepSeek | 1 427 | 1 415–1 438 | 2 686 | MIT |
| 128 | Solar Pro 4 | Upstage | 1 427 | 1 404–1 449 | 682 | Proprietary |
| 129 | Qwen3 VL 235B A22B Thinking | Alibaba Qwen | 1 426 | 1 412–1 441 | 1 614 | Apache 2.0 |
| 130 | Qwen3.5-27B | Alibaba Qwen | 1 425 | 1 418–1 432 | 7 496 | Apache 2.0 |
| 131 | Qwen3 235B A22B Thinking 2507 | Alibaba Qwen | 1 425 | 1 410–1 440 | 1 593 | Apache 2.0 |
| 132 | Gemini 2.5 Flash · flash | 1 423 | 1 419–1 428 | 25 656 | Proprietary | |
| 133 | MiniMax M2.1 | MiniMax | 1 422 | 1 412–1 432 | 3 356 | MIT |
| 134 | Hunyuan Vision 1.5 | Tencent | 1 421 | 1 394–1 449 | 422 | Proprietary |
| 135 | Muse Glimmer 30B | Meta | 1 420 | 1 402–1 439 | 1 051 | Apache-2.0 |
| 136 | Grok 4 Fast · reasoning | xAI | 1 420 | 1 410–1 429 | 3 838 | Proprietary |
| 137 | DeepSeek V3.1 · v3.1-thinking | DeepSeek | 1 418 | 1 404–1 431 | 1 875 | MIT |
| 138 | Qwen3 30B A3B Instruct 2507 | Alibaba Qwen | 1 417 | 1 408–1 426 | 4 600 | Apache 2.0 |
| 139 | MiMo-V2 Flash · flash (thinking) | Xiaomi | 1 416 | 1 404–1 428 | 2 410 | MIT |
| 140 | Grok 4.3 | xAI | 1 414 | 1 408–1 420 | 18 716 | Proprietary |
| 141 | ChatGPT-4o (latest) | OpenAI | 1 414 | 1 409–1 419 | 15 562 | Proprietary |
| 142 | Claude Sonnet 4 · 20250514-thinking-32k | Anthropic | 1 414 | 1 406–1 422 | 6 299 | Proprietary |
| 143 | DeepSeek V3.1 · v3.1 | DeepSeek | 1 414 | 1 402–1 425 | 2 602 | MIT |
| 144 | Amazon Nova Experimental Chat · 10-20 | Amazon | 1 413 | 1 401–1 426 | 2 271 | Proprietary |
| 145 | Qwen3 Coder 480B A35B | Alibaba Qwen | 1 413 | 1 404–1 422 | 4 796 | Apache 2.0 |
| 146 | Qwen3.5-Flash | Alibaba Qwen | 1 413 | 1 407–1 419 | 16 581 | Proprietary |
| 147 | ERNIE 5.0 · preview-1022 | Baidu | 1 413 | 1 394–1 432 | 914 | Proprietary |
| 148 | Grok 4.1 Fast | xAI | 1 411 | 1 405–1 417 | 13 357 | Proprietary |
| 149 | Qwen3.5-35B-A3B | Alibaba Qwen | 1 410 | 1 402–1 417 | 7 962 | Apache 2.0 |
| 150 | Grok 4 | xAI | 1 409 | 1 402–1 415 | 7 928 | Proprietary |
| 151 | o3 | OpenAI | 1 408 | 1 402–1 414 | 11 517 | Proprietary |
| 152 | DeepSeek V3.1 Terminus · terminus | DeepSeek | 1 407 | 1 387–1 428 | 771 | MIT |
| 153 | GPT-5.3 Chat | OpenAI | 1 406 | 1 399–1 413 | 8 698 | Proprietary |
| 154 | GPT-5 Mini в режиме high | OpenAI | 1 406 | 1 397–1 414 | 5 439 | Proprietary |
| 155 | Nemotron 3 Super | NVIDIA | 1 405 | 1 391–1 418 | 1 774 | NVIDIA Open Model |
| 156 | GPT-5.4 Nano в режиме high | OpenAI | 1 404 | 1 398–1 410 | 16 530 | Proprietary |
| 157 | Claude Opus 4 · 20250514 | Anthropic | 1 402 | 1 395–1 409 | 7 785 | Proprietary |
| 158 | Gemini 2.5 Flash · flash-preview-09-2025 | 1 401 | 1 394–1 409 | 6 754 | Proprietary | |
| 159 | GPT-5 | OpenAI | 1 399 | 1 392–1 407 | 5 930 | Proprietary |
| 160 | Gemini 3.1 Flash Lite Preview | 1 399 | 1 393–1 405 | 16 689 | Proprietary | |
| 161 | Kimi K2 0905 | Moonshot AI | 1 399 | 1 386–1 411 | 2 219 | Modified MIT |
| 162 | Qwen3 235B A22B · a22b-no-thinking | Alibaba Qwen | 1 398 | 1 391–1 406 | 6 905 | Apache 2.0 |
| 163 | GLM 4.5 Air | Z.ai | 1 397 | 1 389–1 405 | 6 004 | MIT |
| 164 | GPT-4.5 Preview | OpenAI | 1 397 | 1 384–1 410 | 1 939 | Proprietary |
| 165 | Granite 4.2 30B | IBM Granite | 1 394 | 1 373–1 415 | 818 | Apache 2.0 |
| 166 | Qwen3 Next 80B A3B Thinking | Alibaba Qwen | 1 392 | 1 381–1 404 | 2 642 | Apache 2.0 |
| 167 | GPT-4.1 | OpenAI | 1 391 | 1 385–1 398 | 9 155 | Proprietary |
| 168 | Mercury 2 | Inception Labs | 1 390 | 1 370–1 411 | 758 | Proprietary |
| 169 | Hunyuan T1 | Tencent | 1 390 | 1 370–1 410 | 799 | Proprietary |
| 170 | Ling-flash-2.0 | inclusionAI | 1 390 | 1 375–1 405 | 1 513 | MIT |
| 171 | GLM 4.6V | Z.ai | 1 390 | 1 365–1 415 | 535 | MIT |
| 172 | Nova 2 Lite | Amazon | 1 387 | 1 375–1 399 | 2 499 | Proprietary |
| 173 | Mistral Medium 3 | Mistral AI | 1 386 | 1 378–1 394 | 5 832 | Proprietary |
| 174 | Qwen3 235B A22B · a22b | Alibaba Qwen | 1 385 | 1 376–1 394 | 4 307 | Apache 2.0 |
| 175 | GLM 4.7 Flash | Z.ai | 1 385 | 1 374–1 396 | 2 655 | MIT |
| 176 | gpt-oss-120b | OpenAI | 1 381 | 1 373–1 388 | 6 335 | Apache 2.0 |
| 177 | Claude Sonnet 4 · 20250514 | Anthropic | 1 381 | 1 373–1 388 | 7 273 | Proprietary |
| 178 | Trinity Large | Arcee AI | 1 380 | 1 372–1 387 | 7 569 | Apache 2.0 |
| 179 | Grok 3 Mini в режиме high | xAI | 1 380 | 1 369–1 390 | 3 154 | Proprietary |
| 180 | o3 Mini High | OpenAI | 1 379 | 1 367–1 390 | 2 596 | Proprietary |
| 181 | MiniMax M2.5 | MiniMax | 1 378 | 1 372–1 385 | 10 806 | Modified MIT |
| 182 | Nemotron 3 Nano 30B A3B | NVIDIA | 1 378 | 1 368–1 388 | 3 256 | NVIDIA Open Model |
| 183 | Kimi K2 0711 | Moonshot AI | 1 377 | 1 369–1 386 | 5 179 | Modified MIT |
| 184 | Nemotron 3.5 Lightning | NVIDIA | 1 376 | 1 363–1 389 | 2 235 | OpenMDW-1.1 |
| 185 | Gemini 2.5 Flash Lite · 06-17-thinking | 1 373 | 1 366–1 381 | 5 942 | Proprietary | |
| 186 | Granite 4.2 8B | IBM Granite | 1 373 | 1 351–1 395 | 791 | Apache 2.0 |
| 187 | Gemini 2.5 Flash Lite · 09-2025-no-thinking | 1 372 | 1 365–1 378 | 9 574 | Proprietary | |
| 188 | R1 | DeepSeek | 1 372 | 1 360–1 384 | 2 317 | MIT |
| 189 | MiniMax M2 | MiniMax | 1 371 | 1 356–1 386 | 1 517 | Apache 2.0 |
| 190 | INTELLECT-3 | Prime Intellect | 1 371 | 1 352–1 389 | 971 | MIT |
| 191 | Amazon Nova Experimental Chat · 10-09 | Amazon | 1 370 | 1 346–1 394 | 543 | Proprietary |
| 192 | o4 Mini | OpenAI | 1 369 | 1 362–1 376 | 8 601 | Proprietary |
| 193 | DeepSeek V3 0324 | DeepSeek | 1 369 | 1 362–1 376 | 8 280 | MIT |
| 194 | Step 3 | StepFun | 1 368 | 1 351–1 384 | 1 208 | Apache 2.0 |
| 195 | o1 · 2024-12-17 | OpenAI | 1 367 | 1 358–1 377 | 3 973 | Proprietary |
| 196 | GPT-4.1 Mini | OpenAI | 1 367 | 1 360–1 375 | 6 811 | Proprietary |
| 197 | o1 · preview | OpenAI | 1 367 | 1 358–1 376 | 5 123 | Proprietary |
| 198 | Grok 3 Mini | xAI | 1 366 | 1 357–1 376 | 4 084 | Proprietary |
| 199 | Ring-flash-2.0 | inclusionAI | 1 366 | 1 351–1 381 | 1 490 | MIT |
| 200 | o3 Mini | OpenAI | 1 363 | 1 357–1 370 | 9 385 | Proprietary |
| 201 | Mistral Small 3.2 24B | Mistral AI | 1 363 | 1 353–1 373 | 3 328 | Apache 2.0 |
| 202 | o1-mini | OpenAI | 1 363 | 1 355–1 370 | 8 478 | Proprietary |
| 203 | Granite 4.2 3B | IBM Granite | 1 363 | 1 340–1 385 | 794 | Apache 2.0 |
| 204 | Hunyuan TurboS · 20250416 | Tencent | 1 361 | 1 348–1 375 | 1 758 | Proprietary |
| 205 | Claude 3.7 Sonnet · 20250219-thinking-32k | Anthropic | 1 361 | 1 353–1 369 | 6 130 | Proprietary |
| 206 | Qwen2.5 Max | Alibaba Qwen | 1 360 | 1 351–1 368 | 5 076 | Proprietary |
| 207 | MiniMax M1 | MiniMax | 1 359 | 1 351–1 367 | 6 333 | Apache 2.0 |
| 208 | Trinity Large Thinking | Arcee AI | 1 358 | 1 351–1 366 | 8 116 | Apache 2.0 |
| 209 | Qwen3 32B | Alibaba Qwen | 1 358 | 1 334–1 382 | 513 | Apache 2.0 |
| 210 | Llama 3.3 Nemotron Super 49B v1.5 | NVIDIA | 1 355 | 1 333–1 377 | 653 | Nvidia Open |
| 211 | GPT-5 Nano в режиме high | OpenAI | 1 352 | 1 337–1 366 | 1 662 | Proprietary |
| 212 | Gemini 2.0 Flash | 1 351 | 1 344–1 358 | 6 945 | Proprietary | |
| 213 | GLM 4.5V | Z.ai | 1 349 | 1 330–1 367 | 968 | MIT |
| 214 | OLMo 3.1 32B Instruct | Ai2 | 1 347 | 1 335–1 359 | 2 376 | Apache 2.0 |
| 215 | Hunyuan TurboS · 20250226 | Tencent | 1 343 | 1 312–1 373 | 275 | Proprietary |
| 216 | Claude 3.5 Sonnet · 20241022 | Anthropic | 1 343 | 1 337–1 348 | 14 908 | Proprietary |
| 217 | Step-1o Turbo | StepFun | 1 340 | 1 325–1 355 | 1 484 | Proprietary |
| 218 | Claude 3.7 Sonnet · 20250219 | Anthropic | 1 339 | 1 332–1 346 | 7 055 | Proprietary |
| 219 | Qwen3 30B A3B | Alibaba Qwen | 1 338 | 1 329–1 346 | 4 494 | Apache 2.0 |
| 220 | QwQ 32B | Alibaba Qwen | 1 334 | 1 325–1 343 | 4 010 | Apache 2.0 |
| 221 | Command A | Cohere | 1 331 | 1 324–1 337 | 10 130 | CC-BY-NC-4.0 |
| 222 | Mercury | Inception Labs | 1 329 | 1 301–1 358 | 392 | Proprietary |
| 223 | Qwen-Plus | Alibaba Qwen | 1 328 | 1 310–1 347 | 893 | Proprietary |
| 224 | DeepSeek V3 | DeepSeek | 1 325 | 1 315–1 335 | 3 280 | DeepSeek |
| 225 | Gemma 3 27B | 1 322 | 1 316–1 329 | 7 963 | Gemma | |
| 226 | Gemini 2.0 Flash-Lite | 1 322 | 1 313–1 332 | 3 474 | Proprietary | |
| 227 | OLMo 3 32B Think | Ai2 | 1 321 | 1 303–1 340 | 1 011 | Apache 2.0 |
| 228 | Hunyuan Turbo | Tencent | 1 320 | 1 290–1 349 | 299 | Proprietary |
| 229 | Magistral Medium | Mistral AI | 1 319 | 1 307–1 332 | 2 234 | Proprietary |
| 230 | Step-2 16k | StepFun | 1 317 | 1 297–1 336 | 737 | Proprietary |
| 231 | Qwen2.5 Plus | Alibaba Qwen | 1 315 | 1 301–1 328 | 1 553 | Proprietary |
| 232 | Granite 4.1 8B | IBM Granite | 1 313 | 1 294–1 333 | 1 071 | Apache 2.0 |
| 233 | Yi-Lightning | 01.AI | 1 313 | 1 303–1 322 | 4 316 | Proprietary |
| 234 | Athene V2 Chat | Nexusflow | 1 313 | 1 303–1 322 | 4 019 | NexusFlow |
| 235 | Llama 3.1 Nemotron Ultra 253B v1 | NVIDIA | 1 312 | 1 283–1 342 | 367 | Nvidia Open Model |
| 236 | DeepSeek V2.5 · v2.5-1210 | DeepSeek | 1 310 | 1 293–1 327 | 1 079 | DeepSeek |
| 237 | Mistral Small 3.1 24B | Mistral AI | 1 310 | 1 302–1 318 | 6 081 | Apache 2.0 |
| 238 | gpt-oss-20b | OpenAI | 1 307 | 1 294–1 321 | 2 109 | Apache 2.0 |
| 239 | Hunyuan Large | Tencent | 1 307 | 1 283–1 331 | 519 | Proprietary |
| 240 | Hunyuan Large Vision | Tencent | 1 307 | 1 288–1 326 | 951 | Proprietary |
| 241 | Claude 3.5 Sonnet · 20240620 | Anthropic | 1 307 | 1 300–1 314 | 13 607 | Proprietary |
| 242 | GPT-4.1 Nano | OpenAI | 1 306 | 1 287–1 325 | 807 | Proprietary |
| 243 | Llama 4 Maverick | Meta | 1 302 | 1 295–1 309 | 6 933 | Llama 4 |
| 244 | DeepSeek V2.5 · v2.5 | DeepSeek | 1 302 | 1 292–1 311 | 4 252 | DeepSeek |
| 245 | GPT-4o (2024-05-13) | OpenAI | 1 298 | 1 291–1 304 | 19 526 | Proprietary |
| 246 | Llama 3.3 Nemotron Super 49B v1 | NVIDIA | 1 296 | 1 265–1 327 | 286 | Nvidia |
| 247 | Gemini 1.5 Pro · 1.5-pro-002 | 1 295 | 1 288–1 302 | 9 175 | Proprietary | |
| 248 | Qwen2.5 72B Instruct | Alibaba Qwen | 1 293 | 1 285–1 300 | 6 688 | Qwen |
| 249 | Llama 3.1 405B Instruct · bf16 | Meta | 1 292 | 1 284–1 299 | 6 249 | Llama 3.1 Community |
| 250 | GLM-4-Plus · plus | Z.ai | 1 291 | 1 282–1 300 | 4 449 | Proprietary |
| 251 | OLMo 3.1 32B Think | Ai2 | 1 290 | 1 275–1 306 | 1 512 | Apache 2.0 |
| 252 | GPT-4o-mini (2024-07-18) | OpenAI | 1 290 | 1 284–1 297 | 10 926 | Proprietary |
| 253 | Qwen Max | Alibaba Qwen | 1 288 | 1 277–1 300 | 2 756 | Qwen |
| 254 | Grok 2 | xAI | 1 288 | 1 281–1 295 | 10 368 | Proprietary |
| 255 | GLM-4-Plus · plus-0111 | Z.ai | 1 287 | 1 269–1 305 | 894 | Proprietary |
| 256 | Claude 3.5 Haiku | Anthropic | 1 287 | 1 281–1 293 | 11 131 | Proprietary |
| 257 | Llama 4 Scout | Meta | 1 286 | 1 277–1 294 | 5 208 | Llama |
| 258 | GPT-4o (2024-08-06) | OpenAI | 1 284 | 1 276–1 292 | 7 318 | Proprietary |
| 259 | Llama 3.1 405B Instruct · fp8 | Meta | 1 284 | 1 277–1 291 | 9 714 | Llama 3.1 Community |
| 260 | Gemma 3 12B | 1 282 | 1 259–1 304 | 543 | Gemma | |
| 261 | Mistral Large 2407 | Mistral AI | 1 277 | 1 269–1 285 | 7 589 | Mistral Research |
| 262 | Qwen2.5 Coder 32B Instruct | Alibaba Qwen | 1 276 | 1 258–1 294 | 873 | Apache 2.0 |
| 263 | Mistral Large | Mistral AI | 1 276 | 1 267–1 285 | 4 212 | MRL |
| 264 | Llama 3.1 Nemotron 70B Instruct | NVIDIA | 1 272 | 1 257–1 287 | 1 312 | Llama 3.1 |
| 265 | Nova Pro 1.0 | Amazon | 1 270 | 1 261–1 279 | 3 853 | Proprietary |
| 266 | Hunyuan Standard · 2025-02-10 | Tencent | 1 270 | 1 247–1 294 | 549 | Proprietary |
| 267 | Grok 2 Mini | xAI | 1 269 | 1 262–1 276 | 8 652 | Proprietary |
| 268 | GPT-4 Turbo | OpenAI | 1 269 | 1 262–1 276 | 17 104 | Proprietary |
| 269 | Gemma 3n E4B | 1 269 | 1 258–1 279 | 3 487 | Gemma | |
| 270 | Llama 3.3 70B Instruct | Meta | 1 269 | 1 262–1 275 | 8 718 | Llama-3.3 |
| 271 | Athene 70B | Nexusflow | 1 267 | 1 257–1 278 | 3 122 | CC-BY-NC-4.0 |
| 272 | Gemini 1.5 Pro · 1.5-pro-001 | 1 267 | 1 259–1 274 | 12 747 | Proprietary | |
| 273 | Claude 3 Opus | Anthropic | 1 265 | 1 259–1 270 | 33 748 | Proprietary |
| 274 | Gemini 1.5 Flash · 002 | 1 262 | 1 254–1 270 | 5 892 | Proprietary | |
| 275 | Llama 3.1 70B Instruct | Meta | 1 260 | 1 253–1 267 | 9 389 | Llama 3.1 Community |
| 276 | Gemini 1.5 Pro · advanced-0514 | 1 255 | 1 246–1 264 | 8 138 | Proprietary | |
| 277 | GPT-4 Turbo Preview · 1106-preview | OpenAI | 1 254 | 1 247–1 262 | 15 605 | Proprietary |
| 278 | DeepSeek Coder V2 | DeepSeek | 1 251 | 1 239–1 263 | 2 671 | DeepSeek License |
| 279 | GPT-4 Turbo Preview · 0125-preview | OpenAI | 1 250 | 1 242–1 257 | 15 289 | Proprietary |
| 280 | Granite 4.0 H Small | IBM Granite | 1 248 | 1 231–1 266 | 1 255 | Apache 2.0 |
| 281 | Mistral Small 3 | Mistral AI | 1 246 | 1 234–1 259 | 2 083 | Apache 2.0 |
| 282 | Nova Lite 1.0 | Amazon | 1 239 | 1 229–1 249 | 3 060 | Proprietary |
| 283 | Gemini 1.5 Flash · 001 | 1 236 | 1 228–1 244 | 10 680 | Proprietary | |
| 284 | Llama 3.1 Tulu 3 70B | Ai2 | 1 236 | 1 212–1 259 | 450 | Llama 3.1 |
| 285 | Phi 4 | Microsoft | 1 232 | 1 222–1 242 | 3 305 | MIT |
| 286 | Hunyuan Standard · 256k | Tencent | 1 231 | 1 207–1 255 | 497 | Proprietary |
| 287 | Gemma 3 4B | 1 231 | 1 207–1 254 | 605 | Gemma | |
| 288 | Reka Core | Reka AI | 1 230 | 1 215–1 246 | 1 216 | Proprietary |
| 289 | GLM-4 | Z.ai | 1 228 | 1 214–1 242 | 1 718 | Proprietary |
| 290 | Jamba 1.5 Large | AI21 Labs | 1 228 | 1 214–1 242 | 1 440 | Jamba Open |
| 291 | Llama 3.1 Nemotron 51B Instruct | NVIDIA | 1 223 | 1 202–1 245 | 665 | Llama 3.1 |
| 292 | Claude 3 Sonnet | Anthropic | 1 223 | 1 216–1 230 | 18 888 | Proprietary |
| 293 | Nova Micro 1.0 | Amazon | 1 218 | 1 208–1 228 | 2 981 | Proprietary |
| 294 | Gemini 1.5 Flash-8B | 1 218 | 1 210–1 226 | 6 069 | Proprietary | |
| 295 | Gemma 2 27B | 1 211 | 1 205–1 218 | 12 088 | Gemma license | |
| 296 | OLMo 2 32B Instruct | Ai2 | 1 210 | 1 183–1 238 | 427 | Apache-2.0 |
| 297 | Nemotron-4 340B Instruct | NVIDIA | 1 210 | 1 198–1 221 | 3 254 | NVIDIA Open Model |
| 298 | GPT-4 | OpenAI | 1 209 | 1 200–1 218 | 8 306 | Proprietary |
| 299 | Llama 3 70B Instruct | Meta | 1 207 | 1 200–1 213 | 28 126 | Llama 3 Community |
| 300 | Ministral 8B (2410) | Mistral AI | 1 202 | 1 184–1 221 | 838 | MRL |
Данные на 13 сентября 2026. Как мы считаем: методика
Свежесть данных
Снимок замеров опубликован 13 сентября 2026; всего снимков борда в архиве: 200.
- замеры оценок людей: загружено 18 сентября 2026, 08:25 UTC
- каталог моделей и цены: загружено 18 сентября 2026, 08:51 UTC
- доли использования: загружено 18 сентября 2026, 08:51 UTC
Другие рейтинги
- Рейтинг агентов
- Рейтинг работы с документами
- Рейтинг редактирования изображений
- Рейтинг видео из изображения
- Популярность моделей по использованию через API-агрегаторы
- Рейтинг поисковых моделей
- Рейтинг текстовых моделей
- Рейтинг моделей для творческих текстов
- Рейтинг моделей для математики
- Рейтинг моделей для русского языка
- Рейтинг генерации изображений
- Рейтинг генерации видео
- Рейтинг редактирования видео
- Рейтинг моделей с изображениями на входе
- Рейтинг моделей для веб-разработки
Данные и источники
Обновлено 18 сентября 2026. Каталог, цены и режимы проверяются ежедневно по нескольким API-агрегаторам и провайдерам. Места — по открытым замерам оценок людей, интервалы приведены как в замере.