Рейтинг текстовых моделей
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 «английский», 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 | 1 512 | 1 507–1 516 | 32 019 | Proprietary |
| 2 | Claude Opus 4.6 | Anthropic | 1 506 | 1 501–1 510 | 34 296 | Proprietary |
| 3 | Claude Fable 5.1 в режиме max | Anthropic | 1 506 | 1 494–1 518 | 2 405 | Proprietary |
| 4 | Claude Opus 5 в режиме max | Anthropic | 1 505 | 1 497–1 512 | 8 585 | Proprietary |
| 5 | Gemini 3.8 Flash в режиме high | 1 501 | 1 488–1 515 | 1 986 | Proprietary | |
| 6 | Claude Opus 5 в режиме high | Anthropic | 1 501 | 1 495–1 507 | 17 661 | Proprietary |
| 7 | Claude Fable 5 | Anthropic | 1 499 | 1 493–1 505 | 13 092 | Proprietary |
| 8 | Claude Opus 4.7 в режиме high | Anthropic | 1 495 | 1 490–1 500 | 28 023 | Proprietary |
| 9 | Muse Spark 1.2 в режиме xhigh | Meta | 1 492 | 1 477–1 508 | 1 388 | Proprietary |
| 10 | Qwen3.8 Max (0902) | Alibaba Qwen | 1 490 | 1 482–1 498 | 6 587 | Proprietary |
| 11 | GLM 5.3 в режиме max | Z.ai | 1 489 | 1 480–1 499 | 4 481 | MIT |
| 12 | Gemini 3.7 Flash в режиме high | 1 487 | 1 475–1 499 | 2 365 | Proprietary | |
| 13 | Claude Opus 4.7 | Anthropic | 1 487 | 1 482–1 492 | 28 313 | Proprietary |
| 14 | Muse Spark 1.3 в режиме max | Meta | 1 483 | 1 469–1 496 | 1 907 | Proprietary |
| 15 | Kimi K3 в режиме max | Moonshot AI | 1 480 | 1 473–1 487 | 8 182 | Kimi K3 license |
| 16 | Gemini 3.5 Flash в режиме high | 1 480 | 1 474–1 485 | 16 593 | Proprietary | |
| 17 | Muse Spark | Meta | 1 480 | 1 472–1 487 | 6 468 | Proprietary |
| 18 | Muse Spark 1.1 | Meta | 1 479 | 1 473–1 486 | 11 415 | Proprietary |
| 19 | MiMo-V2.5-Pro | Xiaomi | 1 479 | 1 474–1 484 | 26 164 | MIT |
| 20 | Gemini 3.6 Flash в режиме high | 1 478 | 1 472–1 485 | 11 307 | Proprietary | |
| 21 | Gemini 3 Pro | 1 478 | 1 473–1 483 | 16 443 | Proprietary | |
| 22 | ERNIE 5.1 | Baidu | 1 478 | 1 472–1 484 | 16 985 | Proprietary |
| 23 | Gemini 3.1 Pro Preview | 1 477 | 1 473–1 482 | 47 835 | Proprietary | |
| 24 | Gemini 3.5 Flash в режиме medium | 1 476 | 1 470–1 482 | 15 842 | Proprietary | |
| 25 | Qwen3.5 Max | Alibaba Qwen | 1 475 | 1 469–1 482 | 9 738 | Proprietary |
| 26 | GLM 5.1 | Z.ai | 1 474 | 1 469–1 479 | 21 604 | MIT |
| 27 | GLM 5.3 Flash | Z.ai | 1 474 | 1 465–1 484 | 3 888 | MIT |
| 28 | Claude Sonnet 4.6 | Anthropic | 1 473 | 1 468–1 477 | 29 973 | Proprietary |
| 29 | GLM 5.2 в режиме max | Z.ai | 1 471 | 1 465–1 477 | 15 630 | MIT |
| 30 | Qwen3.7 Max | Alibaba Qwen | 1 471 | 1 457–1 485 | 1 808 | Proprietary |
| 31 | GPT-5.5 в режиме high | OpenAI | 1 469 | 1 464–1 474 | 29 520 | Proprietary |
| 32 | Gemini 3 Flash Preview | 1 468 | 1 462–1 474 | 12 113 | Proprietary | |
| 33 | GPT-5.4 в режиме high | OpenAI | 1 468 | 1 463–1 473 | 27 508 | Proprietary |
| 34 | GPT-5.5 · 5.5 | OpenAI | 1 467 | 1 462–1 472 | 30 236 | Proprietary |
| 35 | Claude Opus 4.8 в режиме high | Anthropic | 1 466 | 1 461–1 471 | 23 236 | Proprietary |
| 36 | Claude Opus 4.5 | Anthropic | 1 460 | 1 456–1 465 | 30 195 | Proprietary |
| 37 | GLM 5 | Z.ai | 1 459 | 1 453–1 465 | 12 305 | MIT |
| 38 | Claude Opus 4.8 | Anthropic | 1 459 | 1 454–1 464 | 23 934 | Proprietary |
| 39 | Qwen3.7 Plus | Alibaba Qwen | 1 459 | 1 453–1 464 | 16 657 | Proprietary |
| 40 | DeepSeek V4 Pro 0423 | DeepSeek | 1 459 | 1 453–1 464 | 23 890 | MIT |
| 41 | Gemini 2.5 Pro | 1 459 | 1 455–1 462 | 52 966 | Proprietary | |
| 42 | Kimi K2.6 | Moonshot AI | 1 458 | 1 453–1 464 | 17 160 | Modified MIT |
| 43 | Amazon Nova Experimental Chat · 26-02-10 | Amazon | 1 458 | 1 442–1 474 | 1 267 | Proprietary |
| 44 | Grok 4.20 · beta-0309-reasoning | xAI | 1 457 | 1 452–1 462 | 28 631 | Proprietary |
| 45 | GPT-6 Astra в режиме max | OpenAI | 1 456 | 1 439–1 474 | 1 139 | Proprietary |
| 46 | Claude Opus 4.5 в режиме high | Anthropic | 1 456 | 1 450–1 462 | 14 459 | Proprietary |
| 47 | GPT-5.6 Sol в режиме xhigh | OpenAI | 1 455 | 1 448–1 461 | 11 162 | Proprietary |
| 48 | Grok 4.5 | xAI | 1 455 | 1 449–1 461 | 12 413 | Proprietary |
| 49 | MiMo-V2 Pro | Xiaomi | 1 454 | 1 448–1 460 | 10 608 | Proprietary |
| 50 | Nemotron 3 Ultra | NVIDIA | 1 454 | 1 445–1 463 | 4 956 | OpenMDW-1.1 |
| 51 | Kimi K2.5 · thinking | Moonshot AI | 1 454 | 1 449–1 458 | 30 097 | Modified MIT |
| 52 | DeepSeek V4 Pro 0423 в режиме high | DeepSeek | 1 454 | 1 449–1 459 | 22 672 | MIT |
| 53 | GLM 4.7 | Z.ai | 1 453 | 1 444–1 462 | 4 887 | MIT |
| 54 | GLM 5V Turbo | Z.ai | 1 453 | 1 443–1 462 | 3 990 | Proprietary |
| 55 | Grok 4.20 Multi-Agent | xAI | 1 452 | 1 447–1 457 | 27 812 | Proprietary |
| 56 | Grok 4.20 · beta1 | xAI | 1 452 | 1 446–1 458 | 12 272 | Proprietary |
| 57 | Claude Sonnet 5 в режиме high | Anthropic | 1 452 | 1 446–1 458 | 15 120 | Proprietary |
| 58 | DeepSeek V4 Pro 0813 в режиме high | DeepSeek | 1 452 | 1 441–1 462 | 3 507 | MIT |
| 59 | GPT-5.4 | OpenAI | 1 451 | 1 446–1 456 | 29 023 | Proprietary |
| 60 | LongCat-Flash-Chat · chat-2602-exp | Meituan | 1 451 | 1 445–1 457 | 13 158 | Proprietary |
| 61 | Gemma 4 31B | 1 451 | 1 439–1 463 | 2 272 | Apache 2.0 | |
| 62 | GPT-5.1 в режиме high | OpenAI | 1 451 | 1 445–1 456 | 16 256 | Proprietary |
| 63 | Claude Sonnet 4.5 | Anthropic | 1 450 | 1 446–1 454 | 34 631 | Proprietary |
| 64 | Qwen3.6 Max Preview | Alibaba Qwen | 1 450 | 1 438–1 462 | 2 385 | Proprietary |
| 65 | Hy3 | Tencent | 1 450 | 1 439–1 460 | 3 121 | Apache 2.0 |
| 66 | Seed 2.0 Pro | ByteDance Seed | 1 449 | 1 445–1 454 | 33 262 | Proprietary |
| 67 | GLM 4.6 | Z.ai | 1 449 | 1 443–1 454 | 15 009 | MIT |
| 68 | Claude Sonnet 4.5 в режиме high | Anthropic | 1 448 | 1 444–1 452 | 35 076 | Proprietary |
| 69 | ERNIE 5.0 · 0110 | Baidu | 1 448 | 1 442–1 453 | 14 566 | Proprietary |
| 70 | Grok 4.1 · 4.1 | xAI | 1 447 | 1 443–1 452 | 28 200 | Proprietary |
| 71 | MiMo-V2.5 | Xiaomi | 1 447 | 1 441–1 453 | 20 177 | MIT |
| 72 | GPT-5.6 Terra в режиме xhigh | OpenAI | 1 446 | 1 440–1 453 | 11 618 | Proprietary |
| 73 | Qwen3.6 Plus | Alibaba Qwen | 1 446 | 1 440–1 451 | 20 498 | Proprietary |
| 74 | GPT-5.2 Chat | OpenAI | 1 445 | 1 440–1 451 | 15 225 | Proprietary |
| 75 | Qwen3.8 27B | Alibaba Qwen | 1 445 | 1 435–1 454 | 4 175 | Apache 2.0 |
| 76 | Gemma 4 26B A4B | 1 445 | 1 432–1 457 | 2 166 | Apache 2.0 | |
| 77 | ERNIE 5.0 · preview-1203 | Baidu | 1 444 | 1 435–1 453 | 3 993 | Proprietary |
| 78 | MiniMax M3 | MiniMax | 1 444 | 1 438–1 449 | 20 621 | MiniMax Community License |
| 79 | Grok 4.1 · 4.1-thinking | xAI | 1 443 | 1 438–1 447 | 27 354 | Proprietary |
| 80 | Qwen3 Max · preview | Alibaba Qwen | 1 442 | 1 436–1 448 | 11 403 | Proprietary |
| 81 | DeepSeek V3.2 Exp · exp-thinking | DeepSeek | 1 442 | 1 432–1 451 | 3 844 | MIT |
| 82 | Inkling | Thinking Machines Lab | 1 442 | 1 435–1 448 | 10 798 | Apache 2.0 |
| 83 | Qwen3.5 397B A17B | Alibaba Qwen | 1 442 | 1 437–1 446 | 34 669 | Apache 2.0 |
| 84 | Amazon Nova Experimental Chat · 12-10 | Amazon | 1 442 | 1 427–1 456 | 1 466 | Proprietary |
| 85 | Gemini 3 Flash Preview в режиме minimal | 1 442 | 1 437–1 446 | 38 178 | Proprietary | |
| 86 | DeepSeek V4 Flash 0423 | DeepSeek | 1 441 | 1 436–1 446 | 21 892 | MIT |
| 87 | DeepSeek V3.2 Exp · exp | DeepSeek | 1 440 | 1 431–1 449 | 4 939 | MIT |
| 88 | Mistral Large 3 2512 | Mistral AI | 1 440 | 1 436–1 444 | 29 042 | Apache 2.0 |
| 89 | DeepSeek V3.2 · v3.2-thinking | DeepSeek | 1 440 | 1 435–1 445 | 16 620 | MIT |
| 90 | LongCat-Flash-Chat · chat | Meituan | 1 440 | 1 431–1 448 | 4 714 | MIT |
| 91 | GPT-5.6 Luna в режиме xhigh | OpenAI | 1 439 | 1 433–1 446 | 11 804 | Proprietary |
| 92 | Gemini 3.5 Flash Lite | 1 439 | 1 432–1 445 | 10 855 | Proprietary | |
| 93 | Mistral Medium 3.1 | Mistral AI | 1 438 | 1 435–1 442 | 39 893 | Proprietary |
| 94 | Grok 3 | xAI | 1 438 | 1 432–1 443 | 17 055 | Proprietary |
| 95 | MiMo-V2 Omni | Xiaomi | 1 438 | 1 430–1 445 | 9 008 | Proprietary |
| 96 | R1 0528 | DeepSeek | 1 437 | 1 430–1 444 | 8 131 | MIT |
| 97 | DeepSeek V3.2 · v3.2 | DeepSeek | 1 437 | 1 432–1 442 | 19 026 | MIT |
| 98 | GLM 4.5 | Z.ai | 1 437 | 1 430–1 443 | 10 069 | MIT |
| 99 | ChatGPT-4o (latest) | OpenAI | 1 435 | 1 431–1 439 | 36 790 | Proprietary |
| 100 | DeepSeek V4 Flash 0423 в режиме high | DeepSeek | 1 435 | 1 429–1 440 | 21 698 | MIT |
| 101 | Kimi K2.5 · instant | Moonshot AI | 1 433 | 1 423–1 444 | 3 123 | Modified MIT |
| 102 | Mistral Medium 3.5 | Mistral AI | 1 433 | 1 425–1 442 | 5 290 | Modified MIT |
| 103 | Grok 4.6 в режиме high | xAI | 1 433 | 1 425–1 441 | 6 413 | Proprietary |
| 104 | DeepSeek V3.1 · v3.1-thinking | DeepSeek | 1 432 | 1 423–1 441 | 4 661 | MIT |
| 105 | Claude Opus 4.1 · 20250805-thinking-16k | Anthropic | 1 432 | 1 427–1 437 | 20 581 | Proprietary |
| 106 | DeepSeek V3.1 · v3.1 | DeepSeek | 1 430 | 1 422–1 438 | 5 993 | MIT |
| 107 | DeepSeek V3.1 Terminus · terminus-thinking | DeepSeek | 1 430 | 1 415–1 445 | 1 456 | MIT |
| 108 | ERNIE 5.0 · preview-1022 | Baidu | 1 430 | 1 418–1 442 | 2 159 | Proprietary |
| 109 | MiMo-V2 Flash · flash (non-thinking) | Xiaomi | 1 430 | 1 425–1 435 | 19 200 | MIT |
| 110 | Qwen3.5-122B-A10B | Alibaba Qwen | 1 430 | 1 424–1 436 | 12 647 | Apache 2.0 |
| 111 | Qwen3 235B A22B Thinking 2507 | Alibaba Qwen | 1 429 | 1 420–1 439 | 3 889 | Apache 2.0 |
| 112 | Qwen3 VL 235B A22B Instruct | Alibaba Qwen | 1 429 | 1 420–1 438 | 4 880 | Apache 2.0 |
| 113 | Hunyuan Vision 1.5 | Tencent | 1 429 | 1 410–1 447 | 955 | Proprietary |
| 114 | Kimi K2 Thinking | Moonshot AI | 1 426 | 1 422–1 431 | 25 468 | Modified MIT |
| 115 | Claude Opus 4.1 · 20250805 | Anthropic | 1 426 | 1 422–1 430 | 31 661 | Proprietary |
| 116 | MiniMax M2.7 | MiniMax | 1 426 | 1 421–1 431 | 29 387 | Modified MIT |
| 117 | GPT-5.1 | OpenAI | 1 425 | 1 420–1 430 | 17 390 | Proprietary |
| 118 | Qwen3 235B A22B Instruct 2507 | Alibaba Qwen | 1 424 | 1 421–1 428 | 40 138 | Apache 2.0 |
| 119 | Qwen3.5-27B | Alibaba Qwen | 1 423 | 1 417–1 429 | 12 225 | Apache 2.0 |
| 120 | Qwen3 Next 80B A3B Instruct | Alibaba Qwen | 1 422 | 1 416–1 428 | 9 632 | Apache 2.0 |
| 121 | GPT-5.2 в режиме high | OpenAI | 1 422 | 1 417–1 427 | 20 448 | Proprietary |
| 122 | Amazon Nova Experimental Chat · 11-10 | Amazon | 1 422 | 1 415–1 428 | 10 206 | Proprietary |
| 123 | Inkling Small | Thinking Machines Lab | 1 421 | 1 414–1 429 | 7 854 | Apache 2.0 |
| 124 | Grok 4.1 Fast | xAI | 1 421 | 1 416–1 425 | 23 602 | Proprietary |
| 125 | Qwen3 VL 235B A22B Thinking | Alibaba Qwen | 1 420 | 1 410–1 430 | 3 455 | Apache 2.0 |
| 126 | GPT-4.5 Preview | OpenAI | 1 420 | 1 413–1 427 | 8 425 | Proprietary |
| 127 | DeepSeek V3.1 Terminus · terminus | DeepSeek | 1 419 | 1 405–1 434 | 1 582 | MIT |
| 128 | Grok 4 | xAI | 1 418 | 1 413–1 424 | 17 390 | Proprietary |
| 129 | Qwen3 Max · 2025-09-23 | Alibaba Qwen | 1 418 | 1 409–1 428 | 3 869 | Proprietary |
| 130 | GPT-5.5 · 5.5-instant | OpenAI | 1 418 | 1 412–1 424 | 12 361 | Proprietary |
| 131 | Gemini 2.5 Flash · flash | 1 418 | 1 415–1 421 | 53 656 | Proprietary | |
| 132 | Step 3.5 Flash | StepFun | 1 418 | 1 413–1 423 | 24 128 | Apache 2.0 |
| 133 | Amazon Nova Experimental Chat · 10-20 | Amazon | 1 417 | 1 408–1 425 | 4 931 | Proprietary |
| 134 | o3 | OpenAI | 1 416 | 1 412–1 421 | 27 116 | Proprietary |
| 135 | Claude Haiku 4.5 | Anthropic | 1 414 | 1 410–1 417 | 56 659 | Proprietary |
| 136 | Gemini 2.5 Flash · flash-preview-09-2025 | 1 413 | 1 408–1 419 | 13 782 | Proprietary | |
| 137 | Hy3 preview | Tencent | 1 413 | 1 403–1 424 | 3 222 | tencent-hunyuan-community |
| 138 | Gemini 3.1 Flash Lite Preview | 1 413 | 1 408–1 418 | 27 871 | Proprietary | |
| 139 | Grok 4 Fast · chat | xAI | 1 413 | 1 401–1 424 | 2 644 | Proprietary |
| 140 | GPT-5.4 Mini в режиме high | OpenAI | 1 412 | 1 407–1 417 | 27 446 | Proprietary |
| 141 | GPT-5.2 | OpenAI | 1 412 | 1 408–1 417 | 35 595 | Proprietary |
| 142 | MiMo-V2 Flash · flash (thinking) | Xiaomi | 1 412 | 1 403–1 421 | 4 225 | MIT |
| 143 | Amazon Nova Experimental Chat · 26-01-10 | Amazon | 1 410 | 1 394–1 425 | 1 349 | Proprietary |
| 144 | GPT-5 в режиме high | OpenAI | 1 409 | 1 403–1 415 | 13 464 | Proprietary |
| 145 | Qwen3.5-35B-A3B | Alibaba Qwen | 1 407 | 1 401–1 413 | 13 043 | Apache 2.0 |
| 146 | GPT-5 | OpenAI | 1 407 | 1 401–1 413 | 13 170 | Proprietary |
| 147 | Qwen3.5-Flash | Alibaba Qwen | 1 405 | 1 400–1 410 | 25 251 | Proprietary |
| 148 | Grok 4 Fast · reasoning | xAI | 1 405 | 1 398–1 412 | 7 971 | Proprietary |
| 149 | Solar Pro 4 | Upstage | 1 405 | 1 386–1 423 | 980 | Proprietary |
| 150 | Grok 4.3 | xAI | 1 403 | 1 399–1 408 | 30 771 | Proprietary |
| 151 | MiniMax M2.1 | MiniMax | 1 403 | 1 395–1 411 | 6 506 | MIT |
| 152 | Nemotron 3 Super | NVIDIA | 1 400 | 1 389–1 410 | 3 031 | NVIDIA Open Model |
| 153 | GLM 4.5 Air | Z.ai | 1 397 | 1 392–1 403 | 13 081 | MIT |
| 154 | Hunyuan T1 | Tencent | 1 397 | 1 383–1 410 | 1 884 | Proprietary |
| 155 | Qwen3 235B A22B · a22b-no-thinking | Alibaba Qwen | 1 397 | 1 391–1 402 | 16 557 | Apache 2.0 |
| 156 | GLM 4.6V | Z.ai | 1 393 | 1 376–1 410 | 1 142 | MIT |
| 157 | Qwen3 Next 80B A3B Thinking | Alibaba Qwen | 1 392 | 1 384–1 400 | 5 855 | Apache 2.0 |
| 158 | GPT-4.1 | OpenAI | 1 392 | 1 387–1 397 | 23 158 | Proprietary |
| 159 | Qwen3 30B A3B Instruct 2507 | Alibaba Qwen | 1 391 | 1 384–1 397 | 9 715 | Apache 2.0 |
| 160 | DeepSeek V3 0324 | DeepSeek | 1 389 | 1 384–1 394 | 21 340 | MIT |
| 161 | GPT-5.3 Chat | OpenAI | 1 389 | 1 383–1 395 | 14 564 | Proprietary |
| 162 | Muse Glimmer 30B | Meta | 1 388 | 1 373–1 404 | 1 506 | Apache-2.0 |
| 163 | Granite 4.2 30B | IBM Granite | 1 386 | 1 371–1 402 | 1 398 | Apache 2.0 |
| 164 | Hunyuan TurboS · 20250416 | Tencent | 1 386 | 1 378–1 394 | 5 393 | Proprietary |
| 165 | Nova 2 Lite | Amazon | 1 386 | 1 377–1 394 | 5 043 | Proprietary |
| 166 | Claude Opus 4 · 20250514-thinking-16k | Anthropic | 1 385 | 1 380–1 391 | 15 394 | Proprietary |
| 167 | Ling-flash-2.0 | inclusionAI | 1 385 | 1 375–1 396 | 2 997 | MIT |
| 168 | R1 | DeepSeek | 1 385 | 1 379–1 391 | 10 721 | MIT |
| 169 | Mistral Medium 3 | Mistral AI | 1 385 | 1 379–1 391 | 15 077 | Proprietary |
| 170 | Amazon Nova Experimental Chat · 10-09 | Amazon | 1 384 | 1 368–1 400 | 1 240 | Proprietary |
| 171 | Gemini 2.5 Flash Lite · 09-2025-no-thinking | 1 383 | 1 379–1 388 | 19 616 | Proprietary | |
| 172 | o1 · preview | OpenAI | 1 383 | 1 376–1 389 | 15 891 | Proprietary |
| 173 | Mercury 2 | Inception Labs | 1 382 | 1 366–1 399 | 1 170 | Proprietary |
| 174 | GPT-5.4 Nano в режиме high | OpenAI | 1 382 | 1 377–1 387 | 27 229 | Proprietary |
| 175 | MiniMax M2.5 | MiniMax | 1 380 | 1 375–1 386 | 18 066 | Modified MIT |
| 176 | INTELLECT-3 | Prime Intellect | 1 380 | 1 368–1 392 | 2 341 | MIT |
| 177 | Grok 3 Mini в режиме high | xAI | 1 379 | 1 372–1 386 | 7 093 | Proprietary |
| 178 | Kimi K2 0905 | Moonshot AI | 1 379 | 1 370–1 387 | 4 764 | Modified MIT |
| 179 | Nemotron 3 Nano 30B A3B | NVIDIA | 1 379 | 1 371–1 386 | 6 306 | NVIDIA Open Model |
| 180 | GPT-5 Mini в режиме high | OpenAI | 1 378 | 1 372–1 384 | 11 353 | Proprietary |
| 181 | Qwen3 235B A22B · a22b | Alibaba Qwen | 1 377 | 1 371–1 383 | 12 346 | Apache 2.0 |
| 182 | Kimi K2 0711 | Moonshot AI | 1 377 | 1 370–1 383 | 11 651 | Modified MIT |
| 183 | Gemma 3 27B | 1 376 | 1 372–1 381 | 22 939 | Gemma | |
| 184 | Claude Opus 4 · 20250514 | Anthropic | 1 376 | 1 370–1 381 | 18 940 | Proprietary |
| 185 | Grok 3 Mini | xAI | 1 376 | 1 369–1 382 | 9 809 | Proprietary |
| 186 | Qwen2.5 Max | Alibaba Qwen | 1 376 | 1 370–1 381 | 18 166 | Proprietary |
| 187 | gpt-oss-120b | OpenAI | 1 375 | 1 369–1 381 | 12 914 | Apache 2.0 |
| 188 | GLM 4.7 Flash | Z.ai | 1 374 | 1 365–1 382 | 4 459 | MIT |
| 189 | Step 3 | StepFun | 1 373 | 1 362–1 384 | 2 767 | Apache 2.0 |
| 190 | Gemini 2.5 Flash Lite · 06-17-thinking | 1 373 | 1 367–1 378 | 13 979 | Proprietary | |
| 191 | o1 · 2024-12-17 | OpenAI | 1 372 | 1 367–1 378 | 16 316 | Proprietary |
| 192 | o4 Mini | OpenAI | 1 368 | 1 363–1 373 | 20 736 | Proprietary |
| 193 | Gemini 2.0 Flash | 1 366 | 1 361–1 370 | 23 750 | Proprietary | |
| 194 | Ring-flash-2.0 | inclusionAI | 1 366 | 1 355–1 376 | 3 014 | MIT |
| 195 | Nemotron 3.5 Lightning | NVIDIA | 1 365 | 1 354–1 375 | 3 467 | OpenMDW-1.1 |
| 196 | MiniMax M2 | MiniMax | 1 365 | 1 354–1 376 | 2 945 | Apache 2.0 |
| 197 | Claude Sonnet 4 · 20250514-thinking-32k | Anthropic | 1 364 | 1 358–1 369 | 14 745 | Proprietary |
| 198 | MiniMax M1 | MiniMax | 1 363 | 1 357–1 369 | 14 819 | Apache 2.0 |
| 199 | GLM 4.5V | Z.ai | 1 363 | 1 350–1 375 | 2 025 | MIT |
| 200 | Qwen3 Coder 480B A35B | Alibaba Qwen | 1 360 | 1 354–1 367 | 10 920 | Apache 2.0 |
| 201 | Qwen3 32B | Alibaba Qwen | 1 359 | 1 347–1 371 | 2 230 | Apache 2.0 |
| 202 | Trinity Large | Arcee AI | 1 358 | 1 352–1 364 | 13 364 | Apache 2.0 |
| 203 | Llama 3.3 Nemotron Super 49B v1.5 | NVIDIA | 1 358 | 1 343–1 373 | 1 379 | Nvidia Open |
| 204 | Mistral Small 3.2 24B | Mistral AI | 1 358 | 1 351–1 365 | 7 518 | Apache 2.0 |
| 205 | Trinity Large Thinking | Arcee AI | 1 357 | 1 351–1 363 | 13 769 | Apache 2.0 |
| 206 | GPT-4.1 Mini | OpenAI | 1 355 | 1 350–1 361 | 18 021 | Proprietary |
| 207 | Llama 3.3 Nemotron Super 49B v1 | NVIDIA | 1 354 | 1 338–1 369 | 1 250 | Nvidia |
| 208 | OLMo 3.1 32B Instruct | Ai2 | 1 353 | 1 344–1 362 | 4 465 | Apache 2.0 |
| 209 | Step-1o Turbo | StepFun | 1 353 | 1 343–1 363 | 3 968 | Proprietary |
| 210 | o3 Mini High | OpenAI | 1 351 | 1 345–1 358 | 11 130 | Proprietary |
| 211 | QwQ 32B | Alibaba Qwen | 1 351 | 1 345–1 356 | 12 890 | Apache 2.0 |
| 212 | Gemma 3 12B | 1 349 | 1 337–1 360 | 2 330 | Gemma | |
| 213 | OLMo 3 32B Think | Ai2 | 1 348 | 1 336–1 360 | 2 398 | Apache 2.0 |
| 214 | DeepSeek V3 | DeepSeek | 1 346 | 1 340–1 352 | 12 866 | DeepSeek |
| 215 | Llama 3.1 Nemotron Ultra 253B v1 | NVIDIA | 1 345 | 1 330–1 360 | 1 491 | Nvidia Open Model |
| 216 | Step-2 16k | StepFun | 1 344 | 1 334–1 355 | 2 781 | Proprietary |
| 217 | Claude Sonnet 4 · 20250514 | Anthropic | 1 344 | 1 338–1 350 | 17 109 | Proprietary |
| 218 | Command A | Cohere | 1 344 | 1 339–1 348 | 26 096 | CC-BY-NC-4.0 |
| 219 | OLMo 3.1 32B Think | Ai2 | 1 342 | 1 332–1 353 | 3 220 | Apache 2.0 |
| 220 | GLM-4-Plus · plus-0111 | Z.ai | 1 340 | 1 330–1 351 | 3 466 | Proprietary |
| 221 | o1-mini | OpenAI | 1 340 | 1 335–1 345 | 27 352 | Proprietary |
| 222 | Qwen-Plus | Alibaba Qwen | 1 339 | 1 328–1 349 | 3 425 | Proprietary |
| 223 | Granite 4.2 8B | IBM Granite | 1 337 | 1 320–1 354 | 1 345 | Apache 2.0 |
| 224 | Qwen3 30B A3B | Alibaba Qwen | 1 335 | 1 329–1 341 | 12 280 | Apache 2.0 |
| 225 | Hunyuan TurboS · 20250226 | Tencent | 1 334 | 1 319–1 349 | 1 310 | Proprietary |
| 226 | Gemini 2.0 Flash-Lite | 1 333 | 1 328–1 338 | 14 891 | Proprietary | |
| 227 | o3 Mini | OpenAI | 1 333 | 1 329–1 337 | 28 604 | Proprietary |
| 228 | Yi-Lightning | 01.AI | 1 330 | 1 323–1 336 | 13 650 | Proprietary |
| 229 | Hunyuan Turbo | Tencent | 1 329 | 1 314–1 343 | 1 319 | Proprietary |
| 230 | Claude 3.7 Sonnet · 20250219-thinking-32k | Anthropic | 1 326 | 1 321–1 332 | 18 864 | Proprietary |
| 231 | Granite 4.2 3B | IBM Granite | 1 326 | 1 309–1 343 | 1 295 | Apache 2.0 |
| 232 | Gemini 1.5 Pro · 1.5-pro-002 | 1 323 | 1 318–1 327 | 30 133 | Proprietary | |
| 233 | Qwen2.5 Plus | Alibaba Qwen | 1 322 | 1 314–1 330 | 5 804 | Proprietary |
| 234 | GPT-5 Nano в режиме high | OpenAI | 1 321 | 1 312–1 331 | 3 575 | Proprietary |
| 235 | Gemma 3n E4B | 1 321 | 1 315–1 328 | 10 065 | Gemma | |
| 236 | Grok 2 | xAI | 1 321 | 1 316–1 325 | 34 362 | Proprietary |
| 237 | Granite 4.1 8B | IBM Granite | 1 317 | 1 303–1 331 | 1 951 | Apache 2.0 |
| 238 | Llama 3.1 Nemotron 70B Instruct | NVIDIA | 1 315 | 1 305–1 325 | 3 714 | Llama 3.1 |
| 239 | GPT-4o (2024-05-13) | OpenAI | 1 312 | 1 308–1 317 | 60 226 | Proprietary |
| 240 | Llama 3.1 405B Instruct · bf16 | Meta | 1 312 | 1 307–1 316 | 23 318 | Llama 3.1 Community |
| 241 | Claude 3.7 Sonnet · 20250219 | Anthropic | 1 310 | 1 305–1 315 | 22 035 | Proprietary |
| 242 | DeepSeek V2.5 · v2.5-1210 | DeepSeek | 1 310 | 1 299–1 320 | 3 776 | DeepSeek |
| 243 | Llama 3.1 405B Instruct · fp8 | Meta | 1 308 | 1 303–1 313 | 32 442 | Llama 3.1 Community |
| 244 | Claude 3.5 Sonnet · 20241022 | Anthropic | 1 307 | 1 303–1 311 | 45 033 | Proprietary |
| 245 | Athene V2 Chat | Nexusflow | 1 307 | 1 301–1 313 | 13 676 | NexusFlow |
| 246 | Hunyuan Large | Tencent | 1 305 | 1 293–1 317 | 2 255 | Proprietary |
| 247 | Llama 4 Scout | Meta | 1 304 | 1 298–1 310 | 12 981 | Llama |
| 248 | Llama 3.3 70B Instruct | Meta | 1 303 | 1 299–1 308 | 29 209 | Llama-3.3 |
| 249 | GPT-4.1 Nano | OpenAI | 1 303 | 1 293–1 313 | 3 528 | Proprietary |
| 250 | gpt-oss-20b | OpenAI | 1 303 | 1 294–1 312 | 4 503 | Apache 2.0 |
| 251 | Llama 4 Maverick | Meta | 1 302 | 1 297–1 308 | 18 862 | Llama 4 |
| 252 | Mercury | Inception Labs | 1 301 | 1 280–1 322 | 841 | Proprietary |
| 253 | Molmo 2 8B | Ai2 | 1 301 | 1 268–1 334 | 325 | Apache 2.0 |
| 254 | GPT-4o-mini (2024-07-18) | OpenAI | 1 301 | 1 296–1 305 | 37 729 | Proprietary |
| 255 | Gemma 3 4B | 1 300 | 1 289–1 312 | 2 572 | Gemma | |
| 256 | Grok 2 Mini | xAI | 1 298 | 1 293–1 303 | 28 429 | Proprietary |
| 257 | GPT-4o (2024-08-06) | OpenAI | 1 297 | 1 292–1 303 | 25 286 | Proprietary |
| 258 | GLM-4-Plus · plus | Z.ai | 1 297 | 1 290–1 303 | 13 191 | Proprietary |
| 259 | Mistral Small 3.1 24B | Mistral AI | 1 295 | 1 289–1 301 | 14 186 | Apache 2.0 |
| 260 | Llama 3.1 70B Instruct | Meta | 1 294 | 1 289–1 298 | 29 694 | Llama 3.1 Community |
| 261 | Qwen Max | Alibaba Qwen | 1 293 | 1 285–1 300 | 8 347 | Qwen |
| 262 | GPT-4 Turbo | OpenAI | 1 291 | 1 286–1 296 | 55 116 | Proprietary |
| 263 | Mistral Large 2407 | Mistral AI | 1 289 | 1 284–1 294 | 24 863 | Mistral Research |
| 264 | Gemini 1.5 Flash · 002 | 1 289 | 1 283–1 294 | 18 364 | Proprietary | |
| 265 | Athene 70B | Nexusflow | 1 288 | 1 281–1 295 | 11 268 | CC-BY-NC-4.0 |
| 266 | Claude 3.5 Sonnet · 20240620 | Anthropic | 1 287 | 1 283–1 292 | 44 433 | Proprietary |
| 267 | DeepSeek V2.5 · v2.5 | DeepSeek | 1 284 | 1 278–1 290 | 12 845 | DeepSeek |
| 268 | Mistral Large | Mistral AI | 1 284 | 1 279–1 289 | 16 289 | MRL |
| 269 | Qwen2.5 72B Instruct | Alibaba Qwen | 1 283 | 1 277–1 288 | 20 561 | Qwen |
| 270 | Gemini 1.5 Pro · advanced-0514 | 1 282 | 1 276–1 288 | 26 318 | Proprietary | |
| 271 | Hunyuan Standard | Tencent | 1 281 | 1 269–1 294 | 2 312 | Proprietary |
| 272 | Hunyuan Large Vision | Tencent | 1 281 | 1 269–1 294 | 2 491 | Proprietary |
| 273 | Gemini 1.5 Pro · 1.5-pro-001 | 1 279 | 1 274–1 284 | 42 538 | Proprietary | |
| 274 | Llama 3 70B Instruct | Meta | 1 277 | 1 273–1 282 | 90 553 | Llama 3 Community |
| 275 | GPT-4 Turbo Preview · 1106-preview | OpenAI | 1 277 | 1 272–1 282 | 64 698 | Proprietary |
| 276 | GPT-4 Turbo Preview · 0125-preview | OpenAI | 1 276 | 1 271–1 281 | 54 967 | Proprietary |
| 277 | Nova Pro 1.0 | Amazon | 1 276 | 1 271–1 282 | 14 188 | Proprietary |
| 278 | Magistral Medium | Mistral AI | 1 274 | 1 265–1 284 | 4 438 | Proprietary |
| 279 | Llama 3.1 Tulu 3 70B | Ai2 | 1 273 | 1 259–1 286 | 1 629 | Llama 3.1 |
| 280 | Claude 3.5 Haiku | Anthropic | 1 267 | 1 263–1 272 | 35 449 | Proprietary |
| 281 | Llama 3.1 Nemotron 51B Instruct | NVIDIA | 1 267 | 1 254–1 281 | 1 794 | Llama 3.1 |
| 282 | Jamba 1.5 Large | AI21 Labs | 1 267 | 1 257–1 276 | 4 915 | Jamba Open |
| 283 | Granite 4.0 H Small | IBM Granite | 1 266 | 1 254–1 279 | 2 434 | Apache 2.0 |
| 284 | Claude 3 Opus | Anthropic | 1 263 | 1 259–1 267 | 108 426 | Proprietary |
| 285 | Mistral Small 3 | Mistral AI | 1 256 | 1 249–1 263 | 8 740 | Apache 2.0 |
| 286 | Reka Core | Reka AI | 1 256 | 1 247–1 265 | 3 934 | Proprietary |
| 287 | OLMo 2 32B Instruct | Ai2 | 1 255 | 1 242–1 268 | 2 025 | Apache-2.0 |
| 288 | Qwen2.5 Coder 32B Instruct | Alibaba Qwen | 1 248 | 1 237–1 259 | 2 652 | Apache 2.0 |
| 289 | Gemini 1.5 Flash · 001 | 1 244 | 1 238–1 249 | 33 810 | Proprietary | |
| 290 | Nova Lite 1.0 | Amazon | 1 243 | 1 236–1 249 | 11 044 | Proprietary |
| 291 | Gemma 2 9B IT SimPO | Princeton NLP | 1 242 | 1 234–1 251 | 5 834 | MIT |
| 292 | GLM-4 | Z.ai | 1 242 | 1 233–1 251 | 5 169 | Proprietary |
| 293 | Gemma 2 27B | 1 242 | 1 238–1 246 | 41 283 | Gemma license | |
| 294 | Command R+ (08-2024) | Cohere | 1 238 | 1 229–1 246 | 5 305 | CC-BY-NC-4.0 |
| 295 | Nemotron-4 340B Instruct | NVIDIA | 1 237 | 1 229–1 244 | 10 292 | NVIDIA Open Model |
| 296 | Gemini 1.5 Flash-8B | 1 231 | 1 226–1 237 | 18 467 | Proprietary | |
| 297 | Phi 4 | Microsoft | 1 231 | 1 225–1 237 | 14 101 | MIT |
| 298 | Aya Expanse 32B | Cohere | 1 229 | 1 223–1 235 | 14 358 | CC-BY-NC-4.0 |
| 299 | Claude 3 Sonnet | Anthropic | 1 226 | 1 221–1 231 | 60 970 | Proprietary |
| 300 | Jamba 1.5 Mini | AI21 Labs | 1 224 | 1 215–1 233 | 5 087 | Jamba Open |
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 моделей для кода
- 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.