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 «сложные запросы», 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 | 1 527 | 1 523–1 532 | 45 675 | Proprietary |
| 2 | Claude Opus 5 в режиме max | Anthropic | 1 524 | 1 518–1 531 | 13 534 | Proprietary |
| 3 | Claude Opus 5 в режиме high | Anthropic | 1 524 | 1 519–1 529 | 28 129 | Proprietary |
| 4 | Claude Opus 4.6 | Anthropic | 1 523 | 1 519–1 527 | 49 172 | Proprietary |
| 5 | Claude Fable 5.1 в режиме max | Anthropic | 1 522 | 1 511–1 532 | 3 437 | Proprietary |
| 6 | Claude Fable 5 | Anthropic | 1 510 | 1 504–1 515 | 19 784 | Proprietary |
| 7 | Gemini 3.8 Flash в режиме high | 1 508 | 1 497–1 518 | 3 426 | Proprietary | |
| 8 | Claude Opus 4.7 в режиме high | Anthropic | 1 506 | 1 501–1 510 | 40 148 | Proprietary |
| 9 | Muse Spark 1.3 в режиме max | Meta | 1 503 | 1 493–1 514 | 3 162 | Proprietary |
| 10 | Claude Opus 4.7 | Anthropic | 1 499 | 1 495–1 504 | 40 875 | Proprietary |
| 11 | Gemini 3.7 Flash в режиме high | 1 496 | 1 486–1 506 | 3 793 | Proprietary | |
| 12 | Qwen3.8 Max (0902) | Alibaba Qwen | 1 495 | 1 488–1 501 | 11 241 | Proprietary |
| 13 | Kimi K3 в режиме max | Moonshot AI | 1 493 | 1 487–1 499 | 13 581 | Kimi K3 license |
| 14 | Muse Spark 1.2 в режиме xhigh | Meta | 1 493 | 1 480–1 506 | 2 115 | Proprietary |
| 15 | GLM 5.3 в режиме max | Z.ai | 1 491 | 1 484–1 499 | 7 161 | MIT |
| 16 | GPT-5.5 в режиме high | OpenAI | 1 489 | 1 484–1 493 | 42 573 | Proprietary |
| 17 | Gemini 3.5 Flash в режиме high | 1 488 | 1 483–1 493 | 25 628 | Proprietary | |
| 18 | GLM 5.3 Flash | Z.ai | 1 488 | 1 480–1 496 | 6 610 | MIT |
| 19 | Muse Spark 1.1 | Meta | 1 488 | 1 482–1 493 | 18 494 | Proprietary |
| 20 | MiMo-V2.5-Pro | Xiaomi | 1 487 | 1 483–1 492 | 39 891 | MIT |
| 21 | GPT-5.4 в режиме high | OpenAI | 1 486 | 1 481–1 491 | 39 229 | Proprietary |
| 22 | Claude Sonnet 4.6 | Anthropic | 1 485 | 1 481–1 490 | 42 987 | Proprietary |
| 23 | Gemini 3.1 Pro Preview | 1 485 | 1 481–1 488 | 69 408 | Proprietary | |
| 24 | Gemini 3.6 Flash в режиме high | 1 484 | 1 479–1 490 | 17 865 | Proprietary | |
| 25 | GPT-5.5 · 5.5 | OpenAI | 1 484 | 1 480–1 489 | 44 063 | Proprietary |
| 26 | Qwen3.5 Max | Alibaba Qwen | 1 483 | 1 477–1 489 | 13 721 | Proprietary |
| 27 | Gemini 3.5 Flash в режиме medium | 1 482 | 1 477–1 487 | 24 642 | Proprietary | |
| 28 | Claude Opus 4.8 в режиме high | Anthropic | 1 482 | 1 477–1 487 | 34 925 | Proprietary |
| 29 | Gemini 3 Pro | 1 482 | 1 477–1 487 | 22 289 | Proprietary | |
| 30 | Qwen3.7 Max | Alibaba Qwen | 1 481 | 1 469–1 494 | 2 524 | Proprietary |
| 31 | ERNIE 5.1 | Baidu | 1 481 | 1 476–1 486 | 23 977 | Proprietary |
| 32 | GPT-5.6 Sol в режиме xhigh | OpenAI | 1 480 | 1 474–1 485 | 17 808 | Proprietary |
| 33 | GLM 5.2 в режиме max | Z.ai | 1 477 | 1 472–1 482 | 24 262 | MIT |
| 34 | Claude Opus 4.5 | Anthropic | 1 476 | 1 472–1 480 | 41 269 | Proprietary |
| 35 | Muse Spark | Meta | 1 474 | 1 467–1 481 | 8 738 | Proprietary |
| 36 | Claude Opus 4.8 | Anthropic | 1 473 | 1 468–1 478 | 35 571 | Proprietary |
| 37 | GLM 5.1 | Z.ai | 1 472 | 1 467–1 476 | 32 274 | MIT |
| 38 | Claude Opus 4.5 в режиме high | Anthropic | 1 472 | 1 467–1 477 | 19 554 | Proprietary |
| 39 | Kimi K2.6 | Moonshot AI | 1 469 | 1 464–1 475 | 24 291 | Modified MIT |
| 40 | GPT-5.4 | OpenAI | 1 469 | 1 464–1 473 | 41 733 | Proprietary |
| 41 | GPT-5.6 Terra в режиме xhigh | OpenAI | 1 468 | 1 462–1 473 | 18 588 | Proprietary |
| 42 | Grok 4.5 | xAI | 1 465 | 1 460–1 471 | 20 145 | Proprietary |
| 43 | Gemini 3 Flash Preview | 1 465 | 1 460–1 471 | 16 475 | Proprietary | |
| 44 | Claude Sonnet 4.5 | Anthropic | 1 462 | 1 459–1 466 | 46 235 | Proprietary |
| 45 | Claude Sonnet 4.5 в режиме high | Anthropic | 1 461 | 1 458–1 465 | 46 520 | Proprietary |
| 46 | DeepSeek V4 Pro 0423 | DeepSeek | 1 461 | 1 456–1 466 | 36 039 | MIT |
| 47 | Claude Sonnet 5 в режиме high | Anthropic | 1 460 | 1 455–1 465 | 23 409 | Proprietary |
| 48 | Qwen3.7 Plus | Alibaba Qwen | 1 459 | 1 454–1 464 | 26 363 | Proprietary |
| 49 | GPT-6 Astra в режиме max | OpenAI | 1 458 | 1 443–1 473 | 1 678 | Proprietary |
| 50 | MiMo-V2 Pro | Xiaomi | 1 458 | 1 452–1 464 | 15 730 | Proprietary |
| 51 | Amazon Nova Experimental Chat · 26-02-10 | Amazon | 1 458 | 1 445–1 471 | 1 923 | Proprietary |
| 52 | DeepSeek V4 Pro 0813 в режиме high | DeepSeek | 1 456 | 1 448–1 464 | 5 967 | MIT |
| 53 | GPT-5.1 в режиме high | OpenAI | 1 456 | 1 451–1 461 | 21 596 | Proprietary |
| 54 | Qwen3.6 Max Preview | Alibaba Qwen | 1 456 | 1 445–1 466 | 3 401 | Proprietary |
| 55 | Qwen3.8 27B | Alibaba Qwen | 1 455 | 1 447–1 463 | 7 078 | Apache 2.0 |
| 56 | Gemini 2.5 Pro | 1 455 | 1 452–1 458 | 64 149 | Proprietary | |
| 57 | Kimi K2.5 · thinking | Moonshot AI | 1 454 | 1 450–1 458 | 44 993 | Modified MIT |
| 58 | Seed 2.0 Pro | ByteDance Seed | 1 453 | 1 449–1 457 | 47 570 | Proprietary |
| 59 | DeepSeek V4 Pro 0423 в режиме high | DeepSeek | 1 452 | 1 448–1 457 | 34 488 | MIT |
| 60 | GLM 5 | Z.ai | 1 452 | 1 447–1 457 | 17 117 | MIT |
| 61 | Nemotron 3 Ultra | NVIDIA | 1 452 | 1 444–1 460 | 7 012 | OpenMDW-1.1 |
| 62 | Grok 4.20 · beta-0309-reasoning | xAI | 1 451 | 1 446–1 455 | 40 552 | Proprietary |
| 63 | MiMo-V2.5 | Xiaomi | 1 450 | 1 445–1 455 | 29 754 | MIT |
| 64 | GPT-5.6 Luna в режиме xhigh | OpenAI | 1 449 | 1 444–1 455 | 18 947 | Proprietary |
| 65 | Hy3 | Tencent | 1 449 | 1 440–1 457 | 5 376 | Apache 2.0 |
| 66 | Qwen3 Max · preview | Alibaba Qwen | 1 449 | 1 443–1 454 | 13 463 | Proprietary |
| 67 | Qwen3.6 Plus | Alibaba Qwen | 1 449 | 1 444–1 453 | 30 478 | Proprietary |
| 68 | Qwen3.5 397B A17B | Alibaba Qwen | 1 448 | 1 445–1 452 | 50 784 | Apache 2.0 |
| 69 | Grok 4.20 Multi-Agent | xAI | 1 448 | 1 443–1 452 | 39 430 | Proprietary |
| 70 | MiniMax M3 | MiniMax | 1 447 | 1 442–1 452 | 32 377 | MiniMax Community License |
| 71 | Inkling | Thinking Machines Lab | 1 447 | 1 441–1 453 | 17 277 | Apache 2.0 |
| 72 | Grok 4.6 в режиме high | xAI | 1 446 | 1 440–1 453 | 10 261 | Proprietary |
| 73 | Gemma 4 31B | 1 446 | 1 436–1 456 | 3 412 | Apache 2.0 | |
| 74 | ERNIE 5.0 · 0110 | Baidu | 1 446 | 1 441–1 450 | 20 778 | Proprietary |
| 75 | GPT-5.2 Chat | OpenAI | 1 445 | 1 440–1 450 | 21 589 | Proprietary |
| 76 | Kimi K2.5 · instant | Moonshot AI | 1 444 | 1 436–1 453 | 4 488 | Modified MIT |
| 77 | MiMo-V2 Omni | Xiaomi | 1 444 | 1 437–1 451 | 12 920 | Proprietary |
| 78 | DeepSeek V4 Flash 0423 | DeepSeek | 1 444 | 1 439–1 449 | 32 523 | MIT |
| 79 | GLM 4.7 | Z.ai | 1 443 | 1 435–1 451 | 6 576 | MIT |
| 80 | GLM 5V Turbo | Z.ai | 1 443 | 1 435–1 451 | 6 180 | Proprietary |
| 81 | Claude Opus 4.1 · 20250805-thinking-16k | Anthropic | 1 443 | 1 438–1 447 | 24 495 | Proprietary |
| 82 | Claude Opus 4.1 · 20250805 | Anthropic | 1 442 | 1 438–1 446 | 38 644 | Proprietary |
| 83 | Gemini 3.5 Flash Lite | 1 441 | 1 435–1 446 | 17 512 | Proprietary | |
| 84 | Grok 4.20 · beta1 | xAI | 1 440 | 1 435–1 446 | 16 601 | Proprietary |
| 85 | GLM 4.6 | Z.ai | 1 440 | 1 435–1 445 | 18 966 | MIT |
| 86 | LongCat-Flash-Chat · chat-2602-exp | Meituan | 1 439 | 1 434–1 445 | 17 793 | Proprietary |
| 87 | ERNIE 5.0 · preview-1203 | Baidu | 1 439 | 1 431–1 448 | 5 151 | Proprietary |
| 88 | Gemma 4 26B A4B | 1 439 | 1 429–1 449 | 3 291 | Apache 2.0 | |
| 89 | Gemini 3 Flash Preview в режиме minimal | 1 438 | 1 434–1 442 | 53 976 | Proprietary | |
| 90 | Mistral Medium 3.5 | Mistral AI | 1 436 | 1 428–1 444 | 7 137 | Modified MIT |
| 91 | Grok 4.1 · 4.1 | xAI | 1 436 | 1 432–1 440 | 37 970 | Proprietary |
| 92 | GPT-5.1 | OpenAI | 1 435 | 1 430–1 439 | 23 477 | Proprietary |
| 93 | Qwen3 235B A22B Instruct 2507 | Alibaba Qwen | 1 434 | 1 431–1 437 | 51 973 | Apache 2.0 |
| 94 | DeepSeek V3.2 · v3.2 | DeepSeek | 1 434 | 1 429–1 438 | 26 034 | MIT |
| 95 | Grok 4.1 · 4.1-thinking | xAI | 1 433 | 1 429–1 437 | 36 923 | Proprietary |
| 96 | Grok 3 | xAI | 1 433 | 1 427–1 440 | 10 577 | Proprietary |
| 97 | LongCat-Flash-Chat · chat | Meituan | 1 433 | 1 425–1 441 | 5 516 | MIT |
| 98 | DeepSeek V4 Flash 0423 в режиме high | DeepSeek | 1 432 | 1 427–1 436 | 32 410 | MIT |
| 99 | DeepSeek V3.2 · v3.2-thinking | DeepSeek | 1 431 | 1 426–1 435 | 22 373 | MIT |
| 100 | GLM 4.5 | Z.ai | 1 429 | 1 423–1 436 | 11 065 | MIT |
| 101 | Qwen3 VL 235B A22B Instruct | Alibaba Qwen | 1 429 | 1 420–1 438 | 5 664 | Apache 2.0 |
| 102 | Amazon Nova Experimental Chat · 12-10 | Amazon | 1 429 | 1 416–1 442 | 1 858 | Proprietary |
| 103 | DeepSeek V3.2 Exp · exp | DeepSeek | 1 429 | 1 421–1 437 | 6 418 | MIT |
| 104 | Mistral Large 3 2512 | Mistral AI | 1 429 | 1 425–1 432 | 41 713 | Apache 2.0 |
| 105 | GPT-5.2 в режиме high | OpenAI | 1 428 | 1 424–1 433 | 28 440 | Proprietary |
| 106 | Kimi K2 Thinking | Moonshot AI | 1 428 | 1 424–1 432 | 35 589 | Modified MIT |
| 107 | Qwen3 Next 80B A3B Instruct | Alibaba Qwen | 1 428 | 1 422–1 434 | 11 831 | Apache 2.0 |
| 108 | DeepSeek V3.2 Exp · exp-thinking | DeepSeek | 1 427 | 1 419–1 436 | 4 669 | MIT |
| 109 | GPT-5.2 | OpenAI | 1 427 | 1 423–1 431 | 49 874 | Proprietary |
| 110 | DeepSeek V3.1 Terminus · terminus-thinking | DeepSeek | 1 427 | 1 412–1 441 | 1 653 | MIT |
| 111 | GPT-5.5 · 5.5-instant | OpenAI | 1 427 | 1 421–1 433 | 17 456 | Proprietary |
| 112 | Mistral Medium 3.1 | Mistral AI | 1 426 | 1 423–1 430 | 51 262 | Proprietary |
| 113 | GPT-5.4 Mini в режиме high | OpenAI | 1 425 | 1 421–1 430 | 38 889 | Proprietary |
| 114 | ChatGPT-4o (latest) | OpenAI | 1 425 | 1 421–1 428 | 37 720 | Proprietary |
| 115 | Qwen3 Max · 2025-09-23 | Alibaba Qwen | 1 424 | 1 415–1 432 | 4 796 | Proprietary |
| 116 | Inkling Small | Thinking Machines Lab | 1 423 | 1 417–1 429 | 12 511 | Apache 2.0 |
| 117 | Gemini 2.5 Flash · flash | 1 422 | 1 419–1 425 | 63 405 | Proprietary | |
| 118 | Amazon Nova Experimental Chat · 11-10 | Amazon | 1 422 | 1 417–1 428 | 13 569 | Proprietary |
| 119 | ERNIE 5.0 · preview-1022 | Baidu | 1 422 | 1 411–1 433 | 2 538 | Proprietary |
| 120 | Qwen3.5-122B-A10B | Alibaba Qwen | 1 422 | 1 417–1 427 | 17 837 | Apache 2.0 |
| 121 | MiniMax M2.7 | MiniMax | 1 422 | 1 418–1 426 | 45 013 | Modified MIT |
| 122 | Claude Haiku 4.5 | Anthropic | 1 421 | 1 417–1 424 | 79 104 | Proprietary |
| 123 | MiMo-V2 Flash · flash (non-thinking) | Xiaomi | 1 420 | 1 415–1 424 | 27 824 | MIT |
| 124 | Hy3 preview | Tencent | 1 419 | 1 409–1 428 | 4 392 | tencent-hunyuan-community |
| 125 | Hunyuan Vision 1.5 | Tencent | 1 418 | 1 401–1 435 | 1 079 | Proprietary |
| 126 | Amazon Nova Experimental Chat · 26-01-10 | Amazon | 1 417 | 1 404–1 430 | 1 837 | Proprietary |
| 127 | DeepSeek V3.1 · v3.1-thinking | DeepSeek | 1 416 | 1 408–1 425 | 5 094 | MIT |
| 128 | GPT-5 в режиме high | OpenAI | 1 416 | 1 410–1 421 | 14 910 | Proprietary |
| 129 | Qwen3 235B A22B Thinking 2507 | Alibaba Qwen | 1 416 | 1 406–1 425 | 3 806 | Apache 2.0 |
| 130 | R1 0528 | DeepSeek | 1 415 | 1 407–1 423 | 6 863 | MIT |
| 131 | DeepSeek V3.1 · v3.1 | DeepSeek | 1 415 | 1 407–1 422 | 6 731 | MIT |
| 132 | Qwen3.5-27B | Alibaba Qwen | 1 413 | 1 408–1 419 | 17 060 | Apache 2.0 |
| 133 | Grok 4 Fast · chat | xAI | 1 412 | 1 401–1 422 | 3 014 | Proprietary |
| 134 | Step 3.5 Flash | StepFun | 1 412 | 1 407–1 416 | 35 344 | Apache 2.0 |
| 135 | MiniMax M2.1 | MiniMax | 1 410 | 1 403–1 416 | 8 941 | MIT |
| 136 | Gemini 2.5 Flash · flash-preview-09-2025 | 1 409 | 1 404–1 415 | 17 366 | Proprietary | |
| 137 | Grok 4 | xAI | 1 409 | 1 404–1 414 | 19 151 | Proprietary |
| 138 | Qwen3 VL 235B A22B Thinking | Alibaba Qwen | 1 407 | 1 398–1 416 | 3 993 | Apache 2.0 |
| 139 | Grok 4.1 Fast | xAI | 1 407 | 1 403–1 411 | 32 283 | Proprietary |
| 140 | Gemini 3.1 Flash Lite Preview | 1 407 | 1 403–1 412 | 38 930 | Proprietary | |
| 141 | DeepSeek V3.1 Terminus · terminus | DeepSeek | 1 406 | 1 393–1 419 | 1 905 | MIT |
| 142 | GPT-5 | OpenAI | 1 405 | 1 400–1 411 | 15 069 | Proprietary |
| 143 | Grok 4 Fast · reasoning | xAI | 1 404 | 1 398–1 411 | 9 667 | Proprietary |
| 144 | Qwen3.5-Flash | Alibaba Qwen | 1 404 | 1 399–1 408 | 37 580 | Proprietary |
| 145 | MiMo-V2 Flash · flash (thinking) | Xiaomi | 1 404 | 1 396–1 412 | 5 932 | MIT |
| 146 | GPT-4.5 Preview | OpenAI | 1 404 | 1 394–1 413 | 3 445 | Proprietary |
| 147 | o3 | OpenAI | 1 402 | 1 397–1 407 | 25 413 | Proprietary |
| 148 | Qwen3.5-35B-A3B | Alibaba Qwen | 1 401 | 1 395–1 406 | 18 332 | Apache 2.0 |
| 149 | Hunyuan T1 | Tencent | 1 399 | 1 386–1 412 | 1 947 | Proprietary |
| 150 | Claude Opus 4 · 20250514-thinking-16k | Anthropic | 1 399 | 1 393–1 405 | 16 131 | Proprietary |
| 151 | Qwen3 30B A3B Instruct 2507 | Alibaba Qwen | 1 398 | 1 391–1 404 | 10 940 | Apache 2.0 |
| 152 | Grok 4.3 | xAI | 1 397 | 1 393–1 402 | 44 555 | Proprietary |
| 153 | GPT-5.3 Chat | OpenAI | 1 397 | 1 392–1 402 | 20 511 | Proprietary |
| 154 | Solar Pro 4 | Upstage | 1 397 | 1 382–1 412 | 1 610 | Proprietary |
| 155 | Amazon Nova Experimental Chat · 10-20 | Amazon | 1 396 | 1 388–1 404 | 5 944 | Proprietary |
| 156 | Muse Glimmer 30B | Meta | 1 396 | 1 383–1 408 | 2 433 | Apache-2.0 |
| 157 | Qwen3 235B A22B · a22b-no-thinking | Alibaba Qwen | 1 394 | 1 388–1 399 | 16 375 | Apache 2.0 |
| 158 | Nemotron 3 Super | NVIDIA | 1 388 | 1 379–1 397 | 4 129 | NVIDIA Open Model |
| 159 | GPT-4.1 | OpenAI | 1 384 | 1 379–1 388 | 21 855 | Proprietary |
| 160 | Kimi K2 0905 | Moonshot AI | 1 382 | 1 374–1 391 | 5 534 | Modified MIT |
| 161 | GPT-5.4 Nano в режиме high | OpenAI | 1 381 | 1 376–1 385 | 38 567 | Proprietary |
| 162 | GPT-5 Mini в режиме high | OpenAI | 1 379 | 1 373–1 385 | 12 707 | Proprietary |
| 163 | GLM 4.5 Air | Z.ai | 1 378 | 1 373–1 384 | 14 713 | MIT |
| 164 | Gemini 2.5 Flash Lite · 09-2025-no-thinking | 1 377 | 1 373–1 382 | 24 866 | Proprietary | |
| 165 | Claude Opus 4 · 20250514 | Anthropic | 1 376 | 1 371–1 382 | 18 857 | Proprietary |
| 166 | Granite 4.2 30B | IBM Granite | 1 375 | 1 361–1 388 | 2 033 | Apache 2.0 |
| 167 | Grok 3 Mini в режиме high | xAI | 1 374 | 1 367–1 381 | 7 242 | Proprietary |
| 168 | Gemini 2.5 Flash Lite · 06-17-thinking | 1 373 | 1 367–1 379 | 14 428 | Proprietary | |
| 169 | Hunyuan TurboS · 20250416 | Tencent | 1 373 | 1 364–1 383 | 3 867 | Proprietary |
| 170 | Qwen3 Coder 480B A35B | Alibaba Qwen | 1 372 | 1 366–1 379 | 11 498 | Apache 2.0 |
| 171 | Claude Sonnet 4 · 20250514-thinking-32k | Anthropic | 1 372 | 1 366–1 377 | 15 339 | Proprietary |
| 172 | o1 · 2024-12-17 | OpenAI | 1 372 | 1 364–1 379 | 6 453 | Proprietary |
| 173 | Qwen3 Next 80B A3B Thinking | Alibaba Qwen | 1 371 | 1 364–1 379 | 6 623 | Apache 2.0 |
| 174 | MiniMax M2.5 | MiniMax | 1 370 | 1 365–1 375 | 25 523 | Modified MIT |
| 175 | GLM 4.6V | Z.ai | 1 367 | 1 351–1 382 | 1 495 | MIT |
| 176 | o3 Mini High | OpenAI | 1 366 | 1 356–1 375 | 4 386 | Proprietary |
| 177 | Kimi K2 0711 | Moonshot AI | 1 365 | 1 359–1 372 | 12 065 | Modified MIT |
| 178 | DeepSeek V3 0324 | DeepSeek | 1 365 | 1 360–1 371 | 18 513 | MIT |
| 179 | Nova 2 Lite | Amazon | 1 365 | 1 357–1 373 | 6 448 | Proprietary |
| 180 | Mistral Medium 3 | Mistral AI | 1 364 | 1 358–1 370 | 13 609 | Proprietary |
| 181 | gpt-oss-120b | OpenAI | 1 364 | 1 358–1 369 | 14 537 | Apache 2.0 |
| 182 | Grok 3 Mini | xAI | 1 363 | 1 356–1 370 | 9 205 | Proprietary |
| 183 | Mercury 2 | Inception Labs | 1 363 | 1 349–1 377 | 1 744 | Proprietary |
| 184 | Ling-flash-2.0 | inclusionAI | 1 363 | 1 353–1 373 | 3 405 | MIT |
| 185 | Qwen3 235B A22B · a22b | Alibaba Qwen | 1 362 | 1 356–1 368 | 10 402 | Apache 2.0 |
| 186 | R1 | DeepSeek | 1 361 | 1 352–1 370 | 4 116 | MIT |
| 187 | Qwen2.5 Max | Alibaba Qwen | 1 360 | 1 353–1 366 | 9 618 | Proprietary |
| 188 | MiniMax M2 | MiniMax | 1 359 | 1 349–1 369 | 3 667 | Apache 2.0 |
| 189 | GLM 4.7 Flash | Z.ai | 1 357 | 1 350–1 365 | 6 439 | MIT |
| 190 | Amazon Nova Experimental Chat · 10-09 | Amazon | 1 356 | 1 341–1 371 | 1 392 | Proprietary |
| 191 | Step 3 | StepFun | 1 355 | 1 345–1 366 | 2 957 | Apache 2.0 |
| 192 | Claude Sonnet 4 · 20250514 | Anthropic | 1 355 | 1 350–1 361 | 17 347 | Proprietary |
| 193 | o1 · preview | OpenAI | 1 354 | 1 347–1 362 | 8 496 | Proprietary |
| 194 | o4 Mini | OpenAI | 1 351 | 1 346–1 356 | 19 278 | Proprietary |
| 195 | INTELLECT-3 | Prime Intellect | 1 350 | 1 339–1 362 | 2 736 | MIT |
| 196 | Trinity Large | Arcee AI | 1 350 | 1 345–1 356 | 18 472 | Apache 2.0 |
| 197 | Nemotron 3 Nano 30B A3B | NVIDIA | 1 349 | 1 342–1 356 | 8 127 | NVIDIA Open Model |
| 198 | GPT-4.1 Mini | OpenAI | 1 348 | 1 343–1 354 | 16 057 | Proprietary |
| 199 | Gemini 2.0 Flash | 1 346 | 1 341–1 352 | 14 031 | Proprietary | |
| 200 | Ring-flash-2.0 | inclusionAI | 1 343 | 1 333–1 353 | 3 487 | MIT |
| 201 | Trinity Large Thinking | Arcee AI | 1 342 | 1 337–1 348 | 19 073 | Apache 2.0 |
| 202 | Nemotron 3.5 Lightning | NVIDIA | 1 342 | 1 333–1 350 | 5 437 | OpenMDW-1.1 |
| 203 | Gemma 3 27B | 1 340 | 1 335–1 345 | 17 611 | Gemma | |
| 204 | MiniMax M1 | MiniMax | 1 339 | 1 334–1 345 | 15 490 | Apache 2.0 |
| 205 | Mistral Small 3.2 24B | Mistral AI | 1 336 | 1 329–1 343 | 7 743 | Apache 2.0 |
| 206 | Step-1o Turbo | StepFun | 1 335 | 1 325–1 345 | 3 680 | Proprietary |
| 207 | GLM 4.5V | Z.ai | 1 335 | 1 323–1 347 | 2 348 | MIT |
| 208 | Qwen3 32B | Alibaba Qwen | 1 334 | 1 318–1 350 | 1 207 | Apache 2.0 |
| 209 | Llama 3.3 Nemotron Super 49B v1.5 | NVIDIA | 1 334 | 1 319–1 349 | 1 429 | Nvidia Open |
| 210 | o1-mini | OpenAI | 1 333 | 1 327–1 339 | 13 899 | Proprietary |
| 211 | Claude 3.7 Sonnet · 20250219-thinking-32k | Anthropic | 1 333 | 1 327–1 339 | 13 694 | Proprietary |
| 212 | o3 Mini | OpenAI | 1 333 | 1 328–1 338 | 19 983 | Proprietary |
| 213 | GPT-5 Nano в режиме high | OpenAI | 1 328 | 1 319–1 338 | 3 780 | Proprietary |
| 214 | Command A | Cohere | 1 326 | 1 321–1 331 | 23 599 | CC-BY-NC-4.0 |
| 215 | QwQ 32B | Alibaba Qwen | 1 325 | 1 319–1 332 | 8 767 | Apache 2.0 |
| 216 | Hunyuan TurboS · 20250226 | Tencent | 1 325 | 1 302–1 347 | 531 | Proprietary |
| 217 | Gemini 2.0 Flash-Lite | 1 324 | 1 317–1 332 | 6 186 | Proprietary | |
| 218 | Granite 4.2 8B | IBM Granite | 1 323 | 1 308–1 337 | 1 974 | Apache 2.0 |
| 219 | OLMo 3.1 32B Instruct | Ai2 | 1 321 | 1 313–1 329 | 6 142 | Apache 2.0 |
| 220 | Qwen-Plus | Alibaba Qwen | 1 317 | 1 302–1 331 | 1 481 | Proprietary |
| 221 | Llama 3.1 Nemotron Ultra 253B v1 | NVIDIA | 1 316 | 1 295–1 337 | 713 | Nvidia Open Model |
| 222 | Claude 3.7 Sonnet · 20250219 | Anthropic | 1 315 | 1 309–1 321 | 15 117 | Proprietary |
| 223 | Qwen3 30B A3B | Alibaba Qwen | 1 315 | 1 308–1 321 | 10 621 | Apache 2.0 |
| 224 | DeepSeek V3 | DeepSeek | 1 312 | 1 304–1 320 | 5 408 | DeepSeek |
| 225 | Llama 3.3 Nemotron Super 49B v1 | NVIDIA | 1 311 | 1 287–1 335 | 510 | Nvidia |
| 226 | Granite 4.2 3B | IBM Granite | 1 310 | 1 295–1 325 | 1 904 | Apache 2.0 |
| 227 | Gemma 3 12B | 1 309 | 1 292–1 327 | 977 | Gemma | |
| 228 | Hunyuan Turbo | Tencent | 1 308 | 1 285–1 331 | 496 | Proprietary |
| 229 | Claude 3.5 Sonnet · 20241022 | Anthropic | 1 305 | 1 301–1 310 | 27 132 | Proprietary |
| 230 | Yi-Lightning | 01.AI | 1 302 | 1 294–1 310 | 6 961 | Proprietary |
| 231 | OLMo 3 32B Think | Ai2 | 1 302 | 1 290–1 313 | 2 932 | Apache 2.0 |
| 232 | Qwen2.5 Plus | Alibaba Qwen | 1 299 | 1 289–1 310 | 2 671 | Proprietary |
| 233 | Step-2 16k | StepFun | 1 299 | 1 284–1 314 | 1 221 | Proprietary |
| 234 | Gemini 1.5 Pro · 1.5-pro-002 | 1 296 | 1 291–1 302 | 14 853 | Proprietary | |
| 235 | Granite 4.1 8B | IBM Granite | 1 294 | 1 282–1 307 | 2 627 | Apache 2.0 |
| 236 | GLM-4-Plus · plus-0111 | Z.ai | 1 294 | 1 279–1 308 | 1 468 | Proprietary |
| 237 | Athene V2 Chat | Nexusflow | 1 293 | 1 285–1 300 | 6 517 | NexusFlow |
| 238 | Molmo 2 8B | Ai2 | 1 292 | 1 265–1 320 | 436 | Apache 2.0 |
| 239 | DeepSeek V2.5 · v2.5-1210 | DeepSeek | 1 289 | 1 276–1 303 | 1 790 | DeepSeek |
| 240 | Mercury | Inception Labs | 1 287 | 1 268–1 305 | 1 026 | Proprietary |
| 241 | Hunyuan Large | Tencent | 1 286 | 1 268–1 305 | 889 | Proprietary |
| 242 | GPT-4.1 Nano | OpenAI | 1 286 | 1 272–1 299 | 1 669 | Proprietary |
| 243 | Gemma 3n E4B | 1 283 | 1 276–1 291 | 8 463 | Gemma | |
| 244 | Llama 4 Maverick | Meta | 1 281 | 1 275–1 287 | 15 851 | Llama 4 |
| 245 | GPT-4o (2024-05-13) | OpenAI | 1 281 | 1 276–1 286 | 32 416 | Proprietary |
| 246 | Mistral Small 3.1 24B | Mistral AI | 1 278 | 1 272–1 284 | 14 571 | Apache 2.0 |
| 247 | Claude 3.5 Sonnet · 20240620 | Anthropic | 1 275 | 1 270–1 281 | 23 249 | Proprietary |
| 248 | gpt-oss-20b | OpenAI | 1 273 | 1 265–1 282 | 4 722 | Apache 2.0 |
| 249 | GLM-4-Plus · plus | Z.ai | 1 273 | 1 265–1 281 | 7 067 | Proprietary |
| 250 | Grok 2 | xAI | 1 272 | 1 267–1 278 | 17 283 | Proprietary |
| 251 | OLMo 3.1 32B Think | Ai2 | 1 272 | 1 262–1 281 | 4 244 | Apache 2.0 |
| 252 | Qwen2.5 72B Instruct | Alibaba Qwen | 1 271 | 1 264–1 277 | 10 545 | Qwen |
| 253 | DeepSeek V2.5 · v2.5 | DeepSeek | 1 270 | 1 263–1 278 | 6 869 | DeepSeek |
| 254 | Qwen Max | Alibaba Qwen | 1 270 | 1 260–1 279 | 4 447 | Qwen |
| 255 | Llama 3.1 405B Instruct · bf16 | Meta | 1 269 | 1 264–1 275 | 10 652 | Llama 3.1 Community |
| 256 | GPT-4o-mini (2024-07-18) | OpenAI | 1 267 | 1 262–1 272 | 18 256 | Proprietary |
| 257 | Magistral Medium | Mistral AI | 1 267 | 1 259–1 275 | 5 618 | Proprietary |
| 258 | Llama 3.1 Nemotron 70B Instruct | NVIDIA | 1 266 | 1 254–1 279 | 1 953 | Llama 3.1 |
| 259 | Llama 4 Scout | Meta | 1 266 | 1 260–1 272 | 12 888 | Llama |
| 260 | Hunyuan Standard · 2025-02-10 | Tencent | 1 264 | 1 246–1 283 | 873 | Proprietary |
| 261 | GPT-4o (2024-08-06) | OpenAI | 1 264 | 1 258–1 270 | 12 655 | Proprietary |
| 262 | Llama 3.1 405B Instruct · fp8 | Meta | 1 264 | 1 258–1 269 | 16 228 | Llama 3.1 Community |
| 263 | Llama 3.3 70B Instruct | Meta | 1 257 | 1 252–1 262 | 16 935 | Llama-3.3 |
| 264 | Gemini 1.5 Flash · 002 | 1 257 | 1 251–1 264 | 9 262 | Proprietary | |
| 265 | Mistral Large | Mistral AI | 1 257 | 1 250–1 264 | 6 954 | MRL |
| 266 | Mistral Large 2407 | Mistral AI | 1 257 | 1 250–1 263 | 12 682 | Mistral Research |
| 267 | Hunyuan Large Vision | Tencent | 1 256 | 1 243–1 270 | 2 147 | Proprietary |
| 268 | Grok 2 Mini | xAI | 1 255 | 1 249–1 261 | 14 208 | Proprietary |
| 269 | Gemini 1.5 Pro · 1.5-pro-001 | 1 255 | 1 249–1 261 | 22 512 | Proprietary | |
| 270 | Gemma 3 4B | 1 253 | 1 235–1 270 | 1 092 | Gemma | |
| 271 | Qwen2.5 Coder 32B Instruct | Alibaba Qwen | 1 251 | 1 237–1 266 | 1 386 | Apache 2.0 |
| 272 | GPT-4 Turbo | OpenAI | 1 251 | 1 245–1 257 | 28 310 | Proprietary |
| 273 | Claude 3.5 Haiku | Anthropic | 1 251 | 1 246–1 255 | 23 518 | Proprietary |
| 274 | Gemini 1.5 Pro · advanced-0514 | 1 248 | 1 241–1 255 | 14 100 | Proprietary | |
| 275 | Nova Pro 1.0 | Amazon | 1 246 | 1 238–1 253 | 6 300 | Proprietary |
| 276 | Claude 3 Opus | Anthropic | 1 246 | 1 241–1 250 | 55 074 | Proprietary |
| 277 | Athene 70B | Nexusflow | 1 243 | 1 235–1 252 | 5 448 | CC-BY-NC-4.0 |
| 278 | GPT-4 Turbo Preview · 1106-preview | OpenAI | 1 241 | 1 235–1 247 | 26 242 | Proprietary |
| 279 | Llama 3.1 70B Instruct | Meta | 1 241 | 1 235–1 247 | 14 981 | Llama 3.1 Community |
| 280 | Granite 4.0 H Small | IBM Granite | 1 239 | 1 228–1 251 | 2 942 | Apache 2.0 |
| 281 | GPT-4 Turbo Preview · 0125-preview | OpenAI | 1 239 | 1 232–1 245 | 25 575 | Proprietary |
| 282 | Mistral Small 3 | Mistral AI | 1 233 | 1 224–1 243 | 3 569 | Apache 2.0 |
| 283 | Llama 3.1 Tulu 3 70B | Ai2 | 1 220 | 1 202–1 239 | 779 | Llama 3.1 |
| 284 | Nova Lite 1.0 | Amazon | 1 220 | 1 212–1 228 | 4 913 | Proprietary |
| 285 | Gemini 1.5 Flash · 001 | 1 220 | 1 213–1 226 | 18 228 | Proprietary | |
| 286 | Phi 4 | Microsoft | 1 220 | 1 212–1 227 | 5 747 | MIT |
| 287 | Hunyuan Standard · 256k | Tencent | 1 218 | 1 198–1 238 | 700 | Proprietary |
| 288 | Jamba 1.5 Large | AI21 Labs | 1 218 | 1 206–1 229 | 2 372 | Jamba Open |
| 289 | GLM-4 | Z.ai | 1 212 | 1 201–1 223 | 2 970 | Proprietary |
| 290 | Reka Core | Reka AI | 1 211 | 1 200–1 223 | 2 145 | Proprietary |
| 291 | Gemini 1.5 Flash-8B | 1 209 | 1 203–1 216 | 9 654 | Proprietary | |
| 292 | OLMo 2 32B Instruct | Ai2 | 1 208 | 1 188–1 227 | 789 | Apache-2.0 |
| 293 | DeepSeek Coder V2 | DeepSeek | 1 207 | 1 197–1 217 | 4 439 | DeepSeek License |
| 294 | Llama 3.1 Nemotron 51B Instruct | NVIDIA | 1 204 | 1 186–1 221 | 1 002 | Llama 3.1 |
| 295 | Nemotron-4 340B Instruct | NVIDIA | 1 202 | 1 193–1 210 | 5 577 | NVIDIA Open Model |
| 296 | GPT-4 | OpenAI | 1 200 | 1 193–1 208 | 13 935 | Proprietary |
| 297 | Gemma 2 27B | 1 198 | 1 193–1 204 | 20 494 | Gemma license | |
| 298 | Claude 3 Sonnet | Anthropic | 1 197 | 1 191–1 203 | 30 725 | Proprietary |
| 299 | Gemma 2 9B IT SimPO | Princeton NLP | 1 196 | 1 185–1 208 | 2 554 | MIT |
| 300 | Llama 3 70B Instruct | Meta | 1 195 | 1 189–1 201 | 45 956 | Llama 3 Community |
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.