Рейтинг моделей для русского языка
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 «русский язык», 300 строк. Оценка и интервал — из источника; интервал не равен диапазону мест.
| Место | Модель | Разработчик | Оценка | Интервал | Голоса | Лицензия |
|---|---|---|---|---|---|---|
| 1 | Claude Fable 5.1 в режиме max | Anthropic | 1 522 | 1 499–1 545 | 715 | Proprietary |
| 2 | Gemini 3.7 Flash в режиме high | 1 511 | 1 486–1 535 | 611 | Proprietary | |
| 3 | Muse Spark 1.2 в режиме xhigh | Meta | 1 508 | 1 476–1 540 | 331 | Proprietary |
| 4 | Claude Opus 5 в режиме max | Anthropic | 1 506 | 1 493–1 519 | 2 264 | Proprietary |
| 5 | Claude Fable 5 | Anthropic | 1 506 | 1 495–1 517 | 3 055 | Proprietary |
| 6 | Claude Opus 5 в режиме high | Anthropic | 1 506 | 1 496–1 515 | 4 656 | Proprietary |
| 7 | Muse Spark 1.3 в режиме max | Meta | 1 501 | 1 476–1 526 | 558 | Proprietary |
| 8 | Claude Opus 4.6 | Anthropic | 1 498 | 1 491–1 506 | 8 207 | Proprietary |
| 9 | Gemini 3.1 Pro Preview | 1 496 | 1 490–1 503 | 11 524 | Proprietary | |
| 10 | Claude Opus 4.7 в режиме high | Anthropic | 1 496 | 1 487–1 504 | 6 063 | Proprietary |
| 11 | Gemini 3.8 Flash в режиме high | 1 495 | 1 472–1 519 | 576 | Proprietary | |
| 12 | Claude Opus 4.6 в режиме high | Anthropic | 1 495 | 1 488–1 503 | 7 796 | Proprietary |
| 13 | Gemini 3 Pro | 1 494 | 1 485–1 503 | 4 791 | Proprietary | |
| 14 | Gemini 3.5 Flash в режиме high | 1 491 | 1 481–1 500 | 4 040 | Proprietary | |
| 15 | Gemini 3.5 Flash в режиме medium | 1 489 | 1 479–1 500 | 3 884 | Proprietary | |
| 16 | Gemini 3.6 Flash в режиме high | 1 487 | 1 476–1 499 | 2 834 | Proprietary | |
| 17 | Claude Opus 4.7 | Anthropic | 1 487 | 1 478–1 495 | 6 201 | Proprietary |
| 18 | Qwen3.8 Max (0902) | Alibaba Qwen | 1 486 | 1 472–1 499 | 1 965 | Proprietary |
| 19 | Muse Spark 1.1 | Meta | 1 484 | 1 473–1 496 | 2 917 | Proprietary |
| 20 | GPT-5.4 в режиме high | OpenAI | 1 483 | 1 475–1 491 | 6 509 | Proprietary |
| 21 | Qwen3.7 Max | Alibaba Qwen | 1 483 | 1 450–1 515 | 374 | Proprietary |
| 22 | Gemini 3 Flash Preview | 1 479 | 1 470–1 489 | 3 914 | Proprietary | |
| 23 | GPT-5.5 в режиме high | OpenAI | 1 475 | 1 467–1 483 | 6 624 | Proprietary |
| 24 | Kimi K3 в режиме max | Moonshot AI | 1 475 | 1 462–1 487 | 2 341 | Kimi K3 license |
| 25 | Claude Opus 4.8 в режиме high | Anthropic | 1 473 | 1 464–1 482 | 5 534 | Proprietary |
| 26 | Qwen3.5 Max | Alibaba Qwen | 1 473 | 1 461–1 485 | 2 395 | Proprietary |
| 27 | GPT-5.5 · 5.5 | OpenAI | 1 473 | 1 465–1 481 | 6 650 | Proprietary |
| 28 | GLM 5.3 Flash | Z.ai | 1 470 | 1 453–1 486 | 1 296 | MIT |
| 29 | Muse Spark | Meta | 1 467 | 1 451–1 483 | 1 468 | Proprietary |
| 30 | GLM 5.3 в режиме max | Z.ai | 1 466 | 1 450–1 483 | 1 295 | MIT |
| 31 | GPT-5.6 Sol в режиме xhigh | OpenAI | 1 465 | 1 454–1 477 | 2 915 | Proprietary |
| 32 | Claude Opus 4.8 | Anthropic | 1 464 | 1 455–1 473 | 5 462 | Proprietary |
| 33 | Gemma 4 31B | 1 461 | 1 441–1 481 | 761 | Apache 2.0 | |
| 34 | Gemini 2.5 Pro | 1 460 | 1 454–1 466 | 11 710 | Proprietary | |
| 35 | GLM 5.2 в режиме max | Z.ai | 1 460 | 1 450–1 470 | 3 910 | MIT |
| 36 | Grok 4.20 · beta-0309-reasoning | xAI | 1 459 | 1 451–1 467 | 6 658 | Proprietary |
| 37 | ERNIE 5.1 | Baidu | 1 458 | 1 448–1 468 | 3 881 | Proprietary |
| 38 | GPT-5.4 | OpenAI | 1 458 | 1 450–1 466 | 6 909 | Proprietary |
| 39 | Qwen3.7 Plus | Alibaba Qwen | 1 458 | 1 448–1 467 | 4 159 | Proprietary |
| 40 | Grok 4.20 Multi-Agent | xAI | 1 458 | 1 450–1 466 | 6 471 | Proprietary |
| 41 | GLM 5.1 | Z.ai | 1 457 | 1 448–1 465 | 5 348 | MIT |
| 42 | Gemini 3 Flash Preview в режиме minimal | 1 453 | 1 446–1 460 | 9 692 | Proprietary | |
| 43 | Claude Sonnet 5 в режиме high | Anthropic | 1 452 | 1 442–1 462 | 3 650 | Proprietary |
| 44 | MiMo-V2.5-Pro | Xiaomi | 1 452 | 1 444–1 460 | 6 744 | MIT |
| 45 | Grok 4.5 | xAI | 1 452 | 1 441–1 463 | 3 289 | Proprietary |
| 46 | GPT-5.6 Terra в режиме xhigh | OpenAI | 1 452 | 1 440–1 463 | 2 950 | Proprietary |
| 47 | DeepSeek V4 Pro 0423 в режиме high | DeepSeek | 1 451 | 1 442–1 459 | 5 527 | MIT |
| 48 | DeepSeek V4 Pro 0423 | DeepSeek | 1 450 | 1 441–1 458 | 5 695 | MIT |
| 49 | Claude Opus 4.5 | Anthropic | 1 448 | 1 441–1 455 | 8 483 | Proprietary |
| 50 | Seed 2.0 Pro | ByteDance Seed | 1 448 | 1 441–1 455 | 8 285 | Proprietary |
| 51 | Qwen3.6 Max Preview | Alibaba Qwen | 1 447 | 1 420–1 473 | 525 | Proprietary |
| 52 | Hy3 | Tencent | 1 446 | 1 426–1 466 | 869 | Apache 2.0 |
| 53 | ERNIE 5.0 · 0110 | Baidu | 1 445 | 1 436–1 454 | 4 264 | Proprietary |
| 54 | Kimi K2.6 | Moonshot AI | 1 445 | 1 435–1 455 | 3 814 | Modified MIT |
| 55 | Grok 4.20 · beta1 | xAI | 1 445 | 1 433–1 456 | 3 038 | Proprietary |
| 56 | Gemini 3.5 Flash Lite | 1 444 | 1 432–1 456 | 2 691 | Proprietary | |
| 57 | DeepSeek V4 Pro 0813 в режиме high | DeepSeek | 1 442 | 1 424–1 460 | 1 066 | MIT |
| 58 | Claude Sonnet 4.6 | Anthropic | 1 440 | 1 433–1 448 | 7 170 | Proprietary |
| 59 | GPT-5.2 Chat | OpenAI | 1 440 | 1 430–1 449 | 4 041 | Proprietary |
| 60 | ERNIE 5.0 · preview-1203 | Baidu | 1 437 | 1 421–1 454 | 1 135 | Proprietary |
| 61 | Gemma 4 26B A4B | 1 437 | 1 418–1 456 | 815 | Apache 2.0 | |
| 62 | GLM 5 | Z.ai | 1 436 | 1 426–1 447 | 3 086 | MIT |
| 63 | GPT-5.1 в режиме high | OpenAI | 1 436 | 1 427–1 445 | 4 904 | Proprietary |
| 64 | DeepSeek V3.1 Terminus · terminus-thinking | DeepSeek | 1 436 | 1 402–1 469 | 286 | MIT |
| 65 | Kimi K2.5 · thinking | Moonshot AI | 1 436 | 1 429–1 443 | 7 992 | Modified MIT |
| 66 | Claude Sonnet 4.5 | Anthropic | 1 435 | 1 429–1 442 | 8 958 | Proprietary |
| 67 | Claude Opus 4.5 в режиме high | Anthropic | 1 435 | 1 426–1 444 | 4 515 | Proprietary |
| 68 | MiniMax M3 | MiniMax | 1 433 | 1 425–1 442 | 5 220 | MiniMax Community License |
| 69 | Qwen3.6 Plus | Alibaba Qwen | 1 433 | 1 424–1 442 | 4 739 | Proprietary |
| 70 | Grok 4.1 · 4.1-thinking | xAI | 1 432 | 1 425–1 440 | 7 463 | Proprietary |
| 71 | Qwen3.5 397B A17B | Alibaba Qwen | 1 430 | 1 423–1 438 | 8 266 | Apache 2.0 |
| 72 | GPT-5.5 · 5.5-instant | OpenAI | 1 430 | 1 418–1 442 | 2 796 | Proprietary |
| 73 | ChatGPT-4o (latest) | OpenAI | 1 429 | 1 422–1 436 | 7 601 | Proprietary |
| 74 | Qwen3 Max · preview | Alibaba Qwen | 1 428 | 1 415–1 442 | 1 834 | Proprietary |
| 75 | DeepSeek V4 Flash 0423 | DeepSeek | 1 427 | 1 418–1 436 | 5 235 | MIT |
| 76 | Grok 4.6 в режиме high | xAI | 1 427 | 1 413–1 441 | 1 802 | Proprietary |
| 77 | Amazon Nova Experimental Chat · 26-02-10 | Amazon | 1 426 | 1 401–1 451 | 450 | Proprietary |
| 78 | Claude Sonnet 4.5 в режиме high | Anthropic | 1 426 | 1 419–1 432 | 8 980 | Proprietary |
| 79 | GPT-5.1 | OpenAI | 1 425 | 1 417–1 434 | 5 113 | Proprietary |
| 80 | Grok 4.1 · 4.1 | xAI | 1 425 | 1 418–1 432 | 7 456 | Proprietary |
| 81 | DeepSeek V4 Flash 0423 в режиме high | DeepSeek | 1 425 | 1 416–1 434 | 5 202 | MIT |
| 82 | GLM 4.7 | Z.ai | 1 424 | 1 410–1 438 | 1 577 | MIT |
| 83 | MiMo-V2 Pro | Xiaomi | 1 424 | 1 413–1 435 | 2 937 | Proprietary |
| 84 | GLM 5V Turbo | Z.ai | 1 424 | 1 405–1 443 | 944 | Proprietary |
| 85 | GPT-5.6 Luna в режиме xhigh | OpenAI | 1 423 | 1 412–1 435 | 2 965 | Proprietary |
| 86 | DeepSeek V3.2 · v3.2 | DeepSeek | 1 423 | 1 415–1 431 | 5 712 | MIT |
| 87 | Inkling | Thinking Machines Lab | 1 422 | 1 411–1 434 | 2 800 | Apache 2.0 |
| 88 | R1 0528 | DeepSeek | 1 422 | 1 407–1 437 | 1 521 | MIT |
| 89 | Qwen3.8 27B | Alibaba Qwen | 1 421 | 1 404–1 438 | 1 210 | Apache 2.0 |
| 90 | Claude Opus 4.1 · 20250805 | Anthropic | 1 421 | 1 413–1 428 | 7 002 | Proprietary |
| 91 | Gemini 3.1 Flash Lite Preview | 1 420 | 1 412–1 428 | 6 541 | Proprietary | |
| 92 | GPT-5.4 Mini в режиме high | OpenAI | 1 420 | 1 411–1 428 | 6 268 | Proprietary |
| 93 | DeepSeek V3.1 Terminus · terminus | DeepSeek | 1 420 | 1 389–1 450 | 309 | MIT |
| 94 | GPT-4.5 Preview | OpenAI | 1 418 | 1 403–1 434 | 1 430 | Proprietary |
| 95 | GLM 4.6 | Z.ai | 1 418 | 1 408–1 428 | 3 408 | MIT |
| 96 | Qwen3 Max · 2025-09-23 | Alibaba Qwen | 1 418 | 1 397–1 439 | 715 | Proprietary |
| 97 | GPT-6 Astra в режиме max | OpenAI | 1 417 | 1 383–1 450 | 329 | Proprietary |
| 98 | Claude Opus 4.1 · 20250805-thinking-16k | Anthropic | 1 416 | 1 407–1 425 | 4 135 | Proprietary |
| 99 | GPT-5.2 | OpenAI | 1 416 | 1 409–1 423 | 8 814 | Proprietary |
| 100 | Grok 3 | xAI | 1 415 | 1 404–1 426 | 2 766 | Proprietary |
| 101 | Nemotron 3 Ultra | NVIDIA | 1 415 | 1 397–1 434 | 1 079 | OpenMDW-1.1 |
| 102 | Gemini 2.5 Flash · flash | 1 415 | 1 410–1 421 | 11 804 | Proprietary | |
| 103 | GLM 4.5 | Z.ai | 1 414 | 1 400–1 428 | 1 578 | MIT |
| 104 | Mistral Large 3 2512 | Mistral AI | 1 413 | 1 406–1 420 | 7 845 | Apache 2.0 |
| 105 | Qwen3 235B A22B Instruct 2507 | Alibaba Qwen | 1 412 | 1 406–1 418 | 9 717 | Apache 2.0 |
| 106 | Mistral Medium 3.1 | Mistral AI | 1 411 | 1 405–1 418 | 9 211 | Proprietary |
| 107 | GPT-5.2 в режиме high | OpenAI | 1 410 | 1 402–1 418 | 5 855 | Proprietary |
| 108 | MiMo-V2 Omni | Xiaomi | 1 409 | 1 396–1 422 | 2 053 | Proprietary |
| 109 | Grok 4 | xAI | 1 409 | 1 398–1 420 | 2 775 | Proprietary |
| 110 | Qwen3 VL 235B A22B Instruct | Alibaba Qwen | 1 408 | 1 388–1 429 | 881 | Apache 2.0 |
| 111 | DeepSeek V3.2 Exp · exp-thinking | DeepSeek | 1 408 | 1 386–1 430 | 626 | MIT |
| 112 | o3 | OpenAI | 1 406 | 1 398–1 415 | 4 644 | Proprietary |
| 113 | Hy3 preview | Tencent | 1 406 | 1 383–1 430 | 651 | tencent-hunyuan-community |
| 114 | GPT-5 | OpenAI | 1 406 | 1 393–1 418 | 2 146 | Proprietary |
| 115 | DeepSeek V3.2 · v3.2-thinking | DeepSeek | 1 405 | 1 397–1 413 | 5 088 | MIT |
| 116 | Kimi K2.5 · instant | Moonshot AI | 1 405 | 1 387–1 422 | 1 030 | Modified MIT |
| 117 | DeepSeek V3.1 · v3.1 | DeepSeek | 1 404 | 1 386–1 423 | 964 | MIT |
| 118 | ERNIE 5.0 · preview-1022 | Baidu | 1 403 | 1 373–1 432 | 343 | Proprietary |
| 119 | Qwen3 Next 80B A3B Instruct | Alibaba Qwen | 1 402 | 1 387–1 416 | 1 571 | Apache 2.0 |
| 120 | Qwen3 235B A22B Thinking 2507 | Alibaba Qwen | 1 401 | 1 378–1 425 | 533 | Apache 2.0 |
| 121 | Grok 4.3 | xAI | 1 401 | 1 393–1 409 | 6 757 | Proprietary |
| 122 | Amazon Nova Experimental Chat · 12-10 | Amazon | 1 401 | 1 375–1 426 | 471 | Proprietary |
| 123 | Mistral Medium 3.5 | Mistral AI | 1 401 | 1 383–1 419 | 1 140 | Modified MIT |
| 124 | DeepSeek V3.2 Exp · exp | DeepSeek | 1 400 | 1 383–1 417 | 1 122 | MIT |
| 125 | GPT-5 в режиме high | OpenAI | 1 400 | 1 387–1 412 | 2 136 | Proprietary |
| 126 | DeepSeek V3.1 · v3.1-thinking | DeepSeek | 1 399 | 1 379–1 419 | 732 | MIT |
| 127 | GPT-5.3 Chat | OpenAI | 1 399 | 1 389–1 409 | 3 750 | Proprietary |
| 128 | Qwen3.5-122B-A10B | Alibaba Qwen | 1 398 | 1 388–1 409 | 3 191 | Apache 2.0 |
| 129 | Amazon Nova Experimental Chat · 11-10 | Amazon | 1 395 | 1 385–1 405 | 3 219 | Proprietary |
| 130 | Gemini 2.5 Flash · flash-preview-09-2025 | 1 395 | 1 384–1 405 | 2 871 | Proprietary | |
| 131 | LongCat-Flash-Chat · chat | Meituan | 1 394 | 1 375–1 414 | 850 | MIT |
| 132 | Qwen3 235B A22B · a22b-no-thinking | Alibaba Qwen | 1 393 | 1 382–1 404 | 2 927 | Apache 2.0 |
| 133 | Muse Glimmer 30B | Meta | 1 393 | 1 364–1 422 | 377 | Apache-2.0 |
| 134 | MiMo-V2.5 | Xiaomi | 1 393 | 1 384–1 402 | 4 760 | MIT |
| 135 | Qwen3.5-27B | Alibaba Qwen | 1 392 | 1 382–1 403 | 3 021 | Apache 2.0 |
| 136 | Claude Opus 4 · 20250514-thinking-16k | Anthropic | 1 392 | 1 380–1 403 | 2 550 | Proprietary |
| 137 | Amazon Nova Experimental Chat · 26-01-10 | Amazon | 1 392 | 1 365–1 419 | 417 | Proprietary |
| 138 | Kimi K2 Thinking | Moonshot AI | 1 391 | 1 384–1 398 | 7 043 | Modified MIT |
| 139 | Grok 4 Fast · reasoning | xAI | 1 389 | 1 373–1 404 | 1 372 | Proprietary |
| 140 | MiniMax M2.1 | MiniMax | 1 388 | 1 377–1 400 | 2 255 | MIT |
| 141 | LongCat-Flash-Chat · chat-2602-exp | Meituan | 1 387 | 1 376–1 398 | 3 004 | Proprietary |
| 142 | MiMo-V2 Flash · flash (non-thinking) | Xiaomi | 1 386 | 1 378–1 394 | 5 801 | MIT |
| 143 | Grok 4.1 Fast | xAI | 1 386 | 1 379–1 393 | 6 673 | Proprietary |
| 144 | Step 3.5 Flash | StepFun | 1 385 | 1 378–1 393 | 6 779 | Apache 2.0 |
| 145 | MiniMax M2.7 | MiniMax | 1 385 | 1 377–1 392 | 7 784 | Modified MIT |
| 146 | Hunyuan T1 | Tencent | 1 384 | 1 353–1 415 | 307 | Proprietary |
| 147 | Kimi K2 0905 | Moonshot AI | 1 383 | 1 364–1 403 | 851 | Modified MIT |
| 148 | Inkling Small | Thinking Machines Lab | 1 383 | 1 369–1 396 | 2 061 | Apache 2.0 |
| 149 | Grok 4 Fast · chat | xAI | 1 381 | 1 356–1 407 | 469 | Proprietary |
| 150 | Qwen3.5-Flash | Alibaba Qwen | 1 381 | 1 373–1 389 | 6 522 | Proprietary |
| 151 | Claude Haiku 4.5 | Anthropic | 1 380 | 1 375–1 386 | 13 751 | Proprietary |
| 152 | Claude Opus 4 · 20250514 | Anthropic | 1 379 | 1 369–1 389 | 3 451 | Proprietary |
| 153 | Qwen3.5-35B-A3B | Alibaba Qwen | 1 378 | 1 368–1 388 | 3 353 | Apache 2.0 |
| 154 | GPT-4.1 | OpenAI | 1 374 | 1 365–1 384 | 3 798 | Proprietary |
| 155 | Gemini 2.5 Flash Lite · 09-2025-no-thinking | 1 373 | 1 364–1 381 | 4 662 | Proprietary | |
| 156 | DeepSeek V3 0324 | DeepSeek | 1 372 | 1 363–1 382 | 3 572 | MIT |
| 157 | GLM 4.5 Air | Z.ai | 1 372 | 1 359–1 385 | 1 972 | MIT |
| 158 | MiMo-V2 Flash · flash (thinking) | Xiaomi | 1 371 | 1 357–1 386 | 1 495 | MIT |
| 159 | Amazon Nova Experimental Chat · 10-20 | Amazon | 1 370 | 1 353–1 388 | 1 092 | Proprietary |
| 160 | Qwen3 30B A3B Instruct 2507 | Alibaba Qwen | 1 369 | 1 354–1 383 | 1 566 | Apache 2.0 |
| 161 | Qwen3 VL 235B A22B Thinking | Alibaba Qwen | 1 367 | 1 345–1 390 | 605 | Apache 2.0 |
| 162 | Hunyuan TurboS · 20250416 | Tencent | 1 367 | 1 349–1 384 | 986 | Proprietary |
| 163 | Qwen3 Coder 480B A35B | Alibaba Qwen | 1 366 | 1 352–1 380 | 1 712 | Apache 2.0 |
| 164 | GPT-5.4 Nano в режиме high | OpenAI | 1 366 | 1 358–1 374 | 6 193 | Proprietary |
| 165 | Gemini 2.5 Flash Lite · 06-17-thinking | 1 362 | 1 350–1 374 | 2 274 | Proprietary | |
| 166 | GPT-5 Mini в режиме high | OpenAI | 1 360 | 1 347–1 373 | 1 842 | Proprietary |
| 167 | Kimi K2 0711 | Moonshot AI | 1 359 | 1 346–1 373 | 1 832 | Modified MIT |
| 168 | Mistral Medium 3 | Mistral AI | 1 358 | 1 347–1 369 | 2 757 | Proprietary |
| 169 | MiniMax M2.5 | MiniMax | 1 356 | 1 347–1 365 | 4 803 | Modified MIT |
| 170 | Solar Pro 4 | Upstage | 1 356 | 1 319–1 392 | 262 | Proprietary |
| 171 | o1 · 2024-12-17 | OpenAI | 1 355 | 1 345–1 366 | 3 078 | Proprietary |
| 172 | Claude Sonnet 4 · 20250514-thinking-32k | Anthropic | 1 354 | 1 342–1 366 | 2 415 | Proprietary |
| 173 | R1 | DeepSeek | 1 354 | 1 341–1 366 | 1 904 | MIT |
| 174 | Claude Sonnet 4 · 20250514 | Anthropic | 1 353 | 1 343–1 364 | 3 197 | Proprietary |
| 175 | Qwen2.5 Max | Alibaba Qwen | 1 353 | 1 342–1 363 | 2 980 | Proprietary |
| 176 | Grok 3 Mini | xAI | 1 352 | 1 338–1 366 | 1 634 | Proprietary |
| 177 | Grok 3 Mini в режиме high | xAI | 1 351 | 1 335–1 368 | 1 149 | Proprietary |
| 178 | Gemini 2.0 Flash | 1 350 | 1 341–1 360 | 3 724 | Proprietary | |
| 179 | Granite 4.2 30B | IBM Granite | 1 350 | 1 317–1 383 | 333 | Apache 2.0 |
| 180 | Gemma 3 27B | 1 348 | 1 339–1 358 | 3 415 | Gemma | |
| 181 | Qwen3 235B A22B · a22b | Alibaba Qwen | 1 343 | 1 331–1 354 | 2 157 | Apache 2.0 |
| 182 | Nova 2 Lite | Amazon | 1 341 | 1 326–1 356 | 1 404 | Proprietary |
| 183 | Trinity Large Thinking | Arcee AI | 1 340 | 1 328–1 351 | 3 072 | Apache 2.0 |
| 184 | Nemotron 3 Super | NVIDIA | 1 340 | 1 321–1 359 | 914 | NVIDIA Open Model |
| 185 | gpt-oss-120b | OpenAI | 1 340 | 1 327–1 352 | 2 120 | Apache 2.0 |
| 186 | INTELLECT-3 | Prime Intellect | 1 339 | 1 314–1 363 | 549 | MIT |
| 187 | GLM 4.6V | Z.ai | 1 339 | 1 309–1 369 | 349 | MIT |
| 188 | Qwen3 Next 80B A3B Thinking | Alibaba Qwen | 1 339 | 1 320–1 357 | 925 | Apache 2.0 |
| 189 | GLM 4.7 Flash | Z.ai | 1 335 | 1 321–1 349 | 1 634 | MIT |
| 190 | Gemma 3 12B | 1 335 | 1 303–1 366 | 285 | Gemma | |
| 191 | MiniMax M2 | MiniMax | 1 334 | 1 308–1 360 | 487 | Apache 2.0 |
| 192 | Llama 3.3 Nemotron Super 49B v1.5 | NVIDIA | 1 332 | 1 296–1 368 | 204 | Nvidia Open |
| 193 | o4 Mini | OpenAI | 1 332 | 1 322–1 341 | 3 422 | Proprietary |
| 194 | Step 3 | StepFun | 1 330 | 1 304–1 355 | 502 | Apache 2.0 |
| 195 | MiniMax M1 | MiniMax | 1 328 | 1 317–1 340 | 2 366 | Apache 2.0 |
| 196 | Gemini 2.0 Flash-Lite | 1 327 | 1 315–1 339 | 2 296 | Proprietary | |
| 197 | Step-1o Turbo | StepFun | 1 326 | 1 305–1 348 | 652 | Proprietary |
| 198 | DeepSeek V3 | DeepSeek | 1 324 | 1 312–1 335 | 2 512 | DeepSeek |
| 199 | Step-2 16k | StepFun | 1 323 | 1 302–1 345 | 642 | Proprietary |
| 200 | Trinity Large | Arcee AI | 1 323 | 1 313–1 334 | 3 476 | Apache 2.0 |
| 201 | Mistral Small 3.2 24B | Mistral AI | 1 323 | 1 307–1 339 | 1 240 | Apache 2.0 |
| 202 | Qwen-Plus | Alibaba Qwen | 1 323 | 1 301–1 345 | 578 | Proprietary |
| 203 | GPT-4.1 Mini | OpenAI | 1 322 | 1 312–1 333 | 3 101 | Proprietary |
| 204 | Gemini 1.5 Pro · 1.5-pro-002 | 1 319 | 1 311–1 327 | 7 738 | Proprietary | |
| 205 | o1 · preview | OpenAI | 1 316 | 1 306–1 325 | 4 488 | Proprietary |
| 206 | Hunyuan TurboS · 20250226 | Tencent | 1 312 | 1 276–1 349 | 217 | Proprietary |
| 207 | Command A | Cohere | 1 312 | 1 304–1 321 | 4 432 | CC-BY-NC-4.0 |
| 208 | GLM-4-Plus · plus-0111 | Z.ai | 1 312 | 1 291–1 334 | 653 | Proprietary |
| 209 | Claude 3.7 Sonnet · 20250219-thinking-32k | Anthropic | 1 310 | 1 300–1 321 | 3 367 | Proprietary |
| 210 | Qwen3 32B | Alibaba Qwen | 1 310 | 1 285–1 335 | 412 | Apache 2.0 |
| 211 | Claude 3.7 Sonnet · 20250219 | Anthropic | 1 310 | 1 300–1 320 | 3 504 | Proprietary |
| 212 | Ling-flash-2.0 | inclusionAI | 1 308 | 1 282–1 335 | 469 | MIT |
| 213 | Nemotron 3 Nano 30B A3B | NVIDIA | 1 305 | 1 292–1 318 | 1 998 | NVIDIA Open Model |
| 214 | Claude 3.5 Sonnet · 20241022 | Anthropic | 1 305 | 1 298–1 312 | 10 320 | Proprietary |
| 215 | Hunyuan Turbo | Tencent | 1 305 | 1 273–1 336 | 266 | Proprietary |
| 216 | Mercury 2 | Inception Labs | 1 304 | 1 276–1 333 | 411 | Proprietary |
| 217 | o3 Mini High | OpenAI | 1 303 | 1 290–1 317 | 1 740 | Proprietary |
| 218 | o3 Mini | OpenAI | 1 301 | 1 292–1 309 | 4 944 | Proprietary |
| 219 | QwQ 32B | Alibaba Qwen | 1 296 | 1 284–1 309 | 2 135 | Apache 2.0 |
| 220 | GLM 4.5V | Z.ai | 1 296 | 1 269–1 323 | 398 | MIT |
| 221 | Gemma 3 4B | 1 293 | 1 263–1 324 | 307 | Gemma | |
| 222 | GPT-5 Nano в режиме high | OpenAI | 1 292 | 1 269–1 316 | 615 | Proprietary |
| 223 | Qwen3 30B A3B | Alibaba Qwen | 1 290 | 1 278–1 302 | 2 237 | Apache 2.0 |
| 224 | DeepSeek V2.5 · v2.5-1210 | DeepSeek | 1 289 | 1 271–1 306 | 1 034 | DeepSeek |
| 225 | Gemma 3n E4B | 1 287 | 1 273–1 301 | 1 644 | Gemma | |
| 226 | Gemini 1.5 Flash · 002 | 1 287 | 1 278–1 296 | 5 101 | Proprietary | |
| 227 | Llama 4 Maverick | Meta | 1 286 | 1 275–1 296 | 3 105 | Llama 4 |
| 228 | Grok 2 | xAI | 1 285 | 1 278–1 293 | 8 745 | Proprietary |
| 229 | GPT-4o (2024-05-13) | OpenAI | 1 285 | 1 278–1 292 | 13 904 | Proprietary |
| 230 | Llama 3.1 Nemotron Ultra 253B v1 | NVIDIA | 1 283 | 1 254–1 312 | 303 | Nvidia Open Model |
| 231 | o1-mini | OpenAI | 1 282 | 1 274–1 290 | 7 413 | Proprietary |
| 232 | Claude 3.5 Sonnet · 20240620 | Anthropic | 1 282 | 1 274–1 290 | 9 966 | Proprietary |
| 233 | Gemini 1.5 Pro · advanced-0514 | 1 282 | 1 271–1 292 | 5 861 | Proprietary | |
| 234 | Athene V2 Chat | Nexusflow | 1 281 | 1 271–1 291 | 3 448 | NexusFlow |
| 235 | Claude 3 Opus | Anthropic | 1 279 | 1 273–1 286 | 21 089 | Proprietary |
| 236 | gpt-oss-20b | OpenAI | 1 277 | 1 255–1 298 | 759 | Apache 2.0 |
| 237 | Granite 4.2 8B | IBM Granite | 1 276 | 1 241–1 311 | 310 | Apache 2.0 |
| 238 | Gemini 1.5 Pro · 1.5-pro-001 | 1 276 | 1 267–1 284 | 9 385 | Proprietary | |
| 239 | GPT-4o-mini (2024-07-18) | OpenAI | 1 275 | 1 267–1 282 | 8 268 | Proprietary |
| 240 | GLM-4-Plus · plus | Z.ai | 1 275 | 1 264–1 285 | 3 904 | Proprietary |
| 241 | Qwen Max | Alibaba Qwen | 1 273 | 1 262–1 285 | 2 432 | Qwen |
| 242 | OLMo 3.1 32B Instruct | Ai2 | 1 273 | 1 258–1 288 | 1 607 | Apache 2.0 |
| 243 | GPT-4o (2024-08-06) | OpenAI | 1 271 | 1 263–1 280 | 5 517 | Proprietary |
| 244 | Qwen2.5 Plus | Alibaba Qwen | 1 270 | 1 255–1 285 | 1 397 | Proprietary |
| 245 | Nemotron 3.5 Lightning | NVIDIA | 1 270 | 1 248–1 291 | 958 | OpenMDW-1.1 |
| 246 | Llama 3.3 Nemotron Super 49B v1 | NVIDIA | 1 269 | 1 234–1 303 | 263 | Nvidia |
| 247 | Hunyuan Large | Tencent | 1 265 | 1 237–1 294 | 374 | Proprietary |
| 248 | Llama 3.1 405B Instruct · fp8 | Meta | 1 265 | 1 257–1 273 | 7 673 | Llama 3.1 Community |
| 249 | Qwen2.5 72B Instruct | Alibaba Qwen | 1 263 | 1 255–1 272 | 5 776 | Qwen |
| 250 | Llama 4 Scout | Meta | 1 263 | 1 251–1 275 | 2 285 | Llama |
| 251 | Mistral Small 3.1 24B | Mistral AI | 1 262 | 1 251–1 274 | 2 419 | Apache 2.0 |
| 252 | Grok 2 Mini | xAI | 1 261 | 1 253–1 269 | 7 193 | Proprietary |
| 253 | GPT-4.1 Nano | OpenAI | 1 260 | 1 240–1 281 | 702 | Proprietary |
| 254 | GPT-4 Turbo | OpenAI | 1 259 | 1 251–1 267 | 9 938 | Proprietary |
| 255 | Hunyuan Standard | Tencent | 1 258 | 1 231–1 285 | 379 | Proprietary |
| 256 | OLMo 3 32B Think | Ai2 | 1 257 | 1 232–1 282 | 586 | Apache 2.0 |
| 257 | Mistral Large 2407 | Mistral AI | 1 256 | 1 248–1 265 | 5 748 | Mistral Research |
| 258 | Llama 3.1 405B Instruct · bf16 | Meta | 1 256 | 1 247–1 264 | 4 863 | Llama 3.1 Community |
| 259 | Yi-Lightning | 01.AI | 1 255 | 1 245–1 265 | 4 189 | Proprietary |
| 260 | Claude 3.5 Haiku | Anthropic | 1 252 | 1 245–1 260 | 6 658 | Proprietary |
| 261 | DeepSeek V2.5 · v2.5 | DeepSeek | 1 252 | 1 242–1 262 | 3 519 | DeepSeek |
| 262 | Llama 3.3 70B Instruct | Meta | 1 251 | 1 243–1 259 | 5 716 | Llama-3.3 |
| 263 | Mistral Large | Mistral AI | 1 251 | 1 241–1 261 | 3 258 | MRL |
| 264 | GPT-4 Turbo Preview · 1106-preview | OpenAI | 1 251 | 1 242–1 260 | 7 053 | Proprietary |
| 265 | Granite 4.2 3B | IBM Granite | 1 250 | 1 214–1 286 | 328 | Apache 2.0 |
| 266 | Athene 70B | Nexusflow | 1 250 | 1 237–1 263 | 1 934 | CC-BY-NC-4.0 |
| 267 | Granite 4.1 8B | IBM Granite | 1 249 | 1 219–1 279 | 438 | Apache 2.0 |
| 268 | Reka Core | Reka AI | 1 249 | 1 231–1 266 | 976 | Proprietary |
| 269 | Llama 3.1 Tulu 3 70B | Ai2 | 1 245 | 1 218–1 272 | 366 | Llama 3.1 |
| 270 | Hunyuan Large Vision | Tencent | 1 242 | 1 215–1 270 | 448 | Proprietary |
| 271 | GPT-4 Turbo Preview · 0125-preview | OpenAI | 1 240 | 1 231–1 249 | 8 356 | Proprietary |
| 272 | Nova Pro 1.0 | Amazon | 1 239 | 1 229–1 250 | 3 014 | Proprietary |
| 273 | Gemini 1.5 Flash · 001 | 1 239 | 1 230–1 248 | 7 468 | Proprietary | |
| 274 | Gemini 1.5 Flash-8B | 1 235 | 1 226–1 244 | 5 317 | Proprietary | |
| 275 | Ring-flash-2.0 | inclusionAI | 1 234 | 1 208–1 260 | 488 | MIT |
| 276 | Gemma 2 27B | 1 233 | 1 226–1 241 | 9 678 | Gemma license | |
| 277 | Llama 3.1 70B Instruct | Meta | 1 233 | 1 225–1 241 | 7 285 | Llama 3.1 Community |
| 278 | Qwen2.5 Coder 32B Instruct | Alibaba Qwen | 1 228 | 1 209–1 246 | 822 | Apache 2.0 |
| 279 | Command R+ (08-2024) | Cohere | 1 227 | 1 212–1 242 | 1 354 | CC-BY-NC-4.0 |
| 280 | Aya Expanse 32B | Cohere | 1 226 | 1 217–1 236 | 3 891 | CC-BY-NC-4.0 |
| 281 | Llama 3.1 Nemotron 70B Instruct | NVIDIA | 1 226 | 1 209–1 244 | 1 069 | Llama 3.1 |
| 282 | Claude 3 Sonnet | Anthropic | 1 226 | 1 217–1 235 | 10 794 | Proprietary |
| 283 | Gemma 2 9B IT SimPO | Princeton NLP | 1 224 | 1 207–1 241 | 1 119 | MIT |
| 284 | Magistral Medium | Mistral AI | 1 223 | 1 204–1 243 | 929 | Proprietary |
| 285 | Nemotron-4 340B Instruct | NVIDIA | 1 219 | 1 206–1 231 | 2 494 | NVIDIA Open Model |
| 286 | Mistral Small 3 | Mistral AI | 1 215 | 1 201–1 230 | 1 418 | Apache 2.0 |
| 287 | Nova Lite 1.0 | Amazon | 1 215 | 1 204–1 227 | 2 470 | Proprietary |
| 288 | GLM-4 | Z.ai | 1 213 | 1 196–1 229 | 1 264 | Proprietary |
| 289 | Reka Flash (2024-09) | Reka AI | 1 211 | 1 194–1 228 | 1 012 | Proprietary |
| 290 | Phi 4 | Microsoft | 1 208 | 1 197–1 219 | 2 576 | MIT |
| 291 | Command R+ | Cohere | 1 205 | 1 196–1 215 | 7 778 | CC-BY-NC-4.0 |
| 292 | Claude 3 Haiku | Anthropic | 1 204 | 1 195–1 212 | 11 956 | Proprietary |
| 293 | Gemma 2 9B | 1 200 | 1 192–1 208 | 6 684 | Gemma license | |
| 294 | Granite 4.0 H Small | IBM Granite | 1 197 | 1 168–1 226 | 417 | Apache 2.0 |
| 295 | Aya Expanse 8B | Cohere | 1 197 | 1 182–1 211 | 1 474 | CC-BY-NC-4.0 |
| 296 | Jamba 1.5 Large | AI21 Labs | 1 195 | 1 177–1 213 | 998 | Jamba Open |
| 297 | GPT-4 | OpenAI | 1 195 | 1 183–1 207 | 3 579 | Proprietary |
| 298 | Ministral 8B (2410) | Mistral AI | 1 195 | 1 175–1 214 | 760 | MRL |
| 299 | Llama 3.1 Tulu 3 8B | Ai2 | 1 193 | 1 168–1 217 | 418 | Llama 3.1 |
| 300 | OLMo 3.1 32B Think | Ai2 | 1 192 | 1 173–1 211 | 1 037 | Apache 2.0 |
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 кода
- Рейтинг моделей для творческих текстов
- Рейтинг моделей для математики
- 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.