Рейтинг моделей для творческих текстов
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 «творческие тексты», 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 | 1 508 | 1 490–1 526 | 1 233 | Proprietary |
| 2 | Claude Opus 4.6 в режиме high | Anthropic | 1 505 | 1 498–1 511 | 12 646 | Proprietary |
| 3 | Gemini 3.8 Flash в режиме high | 1 498 | 1 479–1 517 | 1 072 | Proprietary | |
| 4 | Claude Fable 5 | Anthropic | 1 496 | 1 488–1 505 | 6 073 | Proprietary |
| 5 | Gemini 3.7 Flash в режиме high | 1 492 | 1 473–1 510 | 1 185 | Proprietary | |
| 6 | Claude Opus 5 в режиме max | Anthropic | 1 492 | 1 482–1 501 | 4 529 | Proprietary |
| 7 | Claude Opus 5 в режиме high | Anthropic | 1 488 | 1 481–1 496 | 9 196 | Proprietary |
| 8 | Claude Opus 4.7 в режиме high | Anthropic | 1 485 | 1 478–1 492 | 10 511 | Proprietary |
| 9 | Claude Opus 4.6 | Anthropic | 1 484 | 1 478–1 491 | 12 704 | Proprietary |
| 10 | Gemini 3.1 Pro Preview | 1 482 | 1 476–1 487 | 18 759 | Proprietary | |
| 11 | Qwen3.8 Max (0902) | Alibaba Qwen | 1 481 | 1 470–1 492 | 3 232 | Proprietary |
| 12 | Gemini 3 Pro | 1 481 | 1 472–1 489 | 6 237 | Proprietary | |
| 13 | Claude Opus 4.7 | Anthropic | 1 478 | 1 471–1 485 | 10 755 | Proprietary |
| 14 | Gemini 3.5 Flash в режиме high | 1 469 | 1 461–1 477 | 7 107 | Proprietary | |
| 15 | Gemini 3.5 Flash в режиме medium | 1 468 | 1 459–1 476 | 6 997 | Proprietary | |
| 16 | Qwen3.5 Max | Alibaba Qwen | 1 466 | 1 455–1 477 | 3 150 | Proprietary |
| 17 | Gemini 3.6 Flash в режиме high | 1 464 | 1 455–1 473 | 5 352 | Proprietary | |
| 18 | GLM 5.3 в режиме max | Z.ai | 1 462 | 1 449–1 475 | 2 436 | MIT |
| 19 | GLM 5.2 в режиме max | Z.ai | 1 461 | 1 453–1 469 | 7 046 | MIT |
| 20 | Gemini 3 Flash Preview | 1 458 | 1 449–1 468 | 4 619 | Proprietary | |
| 21 | Muse Spark | Meta | 1 458 | 1 444–1 472 | 1 948 | Proprietary |
| 22 | GPT-5.5 в режиме high | OpenAI | 1 454 | 1 447–1 461 | 11 868 | Proprietary |
| 23 | Gemini 2.5 Pro | 1 453 | 1 448–1 458 | 17 131 | Proprietary | |
| 24 | GPT-5.6 Sol в режиме xhigh | OpenAI | 1 452 | 1 443–1 461 | 5 344 | Proprietary |
| 25 | Muse Spark 1.3 в режиме max | Meta | 1 452 | 1 432–1 473 | 902 | Proprietary |
| 26 | GLM 5.1 | Z.ai | 1 452 | 1 445–1 459 | 9 114 | MIT |
| 27 | Kimi K3 в режиме max | Moonshot AI | 1 452 | 1 442–1 462 | 4 133 | Kimi K3 license |
| 28 | Qwen3.7 Max | Alibaba Qwen | 1 451 | 1 424–1 479 | 479 | Proprietary |
| 29 | GPT-5.5 · 5.5 | OpenAI | 1 451 | 1 444–1 458 | 12 171 | Proprietary |
| 30 | Claude Opus 4.8 в режиме high | Anthropic | 1 450 | 1 443–1 457 | 10 012 | Proprietary |
| 31 | Muse Spark 1.2 в режиме xhigh | Meta | 1 448 | 1 423–1 472 | 615 | Proprietary |
| 32 | Claude Opus 4.8 | Anthropic | 1 446 | 1 439–1 454 | 9 982 | Proprietary |
| 33 | Claude Opus 4.5 в режиме high | Anthropic | 1 444 | 1 435–1 452 | 5 456 | Proprietary |
| 34 | DeepSeek V4 Pro 0423 | DeepSeek | 1 444 | 1 436–1 451 | 8 998 | MIT |
| 35 | Muse Spark 1.1 | Meta | 1 443 | 1 434–1 451 | 5 458 | Proprietary |
| 36 | Claude Opus 4.5 | Anthropic | 1 442 | 1 435–1 448 | 10 833 | Proprietary |
| 37 | GLM 5.3 Flash | Z.ai | 1 441 | 1 428–1 455 | 2 138 | MIT |
| 38 | ERNIE 5.1 | Baidu | 1 441 | 1 433–1 450 | 5 907 | Proprietary |
| 39 | Claude Sonnet 4.5 | Anthropic | 1 441 | 1 435–1 447 | 11 841 | Proprietary |
| 40 | DeepSeek V4 Pro 0423 в режиме high | DeepSeek | 1 441 | 1 433–1 448 | 8 621 | MIT |
| 41 | Grok 4.5 | xAI | 1 440 | 1 432–1 449 | 5 946 | Proprietary |
| 42 | GPT-5.4 в режиме high | OpenAI | 1 440 | 1 433–1 447 | 10 081 | Proprietary |
| 43 | MiMo-V2.5-Pro | Xiaomi | 1 439 | 1 433–1 446 | 10 955 | MIT |
| 44 | Qwen3.7 Plus | Alibaba Qwen | 1 439 | 1 431–1 447 | 7 224 | Proprietary |
| 45 | GLM 5 | Z.ai | 1 438 | 1 428–1 447 | 4 490 | MIT |
| 46 | Grok 4.20 · beta1 | xAI | 1 437 | 1 427–1 447 | 3 946 | Proprietary |
| 47 | Grok 4.6 в режиме high | xAI | 1 437 | 1 425–1 448 | 3 222 | Proprietary |
| 48 | Grok 4.20 Multi-Agent | xAI | 1 437 | 1 430–1 444 | 10 210 | Proprietary |
| 49 | Claude Sonnet 4.6 | Anthropic | 1 436 | 1 430–1 443 | 11 259 | Proprietary |
| 50 | Qwen3.6 Max Preview | Alibaba Qwen | 1 435 | 1 412–1 458 | 717 | Proprietary |
| 51 | DeepSeek V4 Pro 0813 в режиме high | DeepSeek | 1 435 | 1 420–1 449 | 1 834 | MIT |
| 52 | Gemini 3 Flash Preview в режиме minimal | 1 434 | 1 428–1 440 | 13 964 | Proprietary | |
| 53 | Grok 4.20 · beta-0309-reasoning | xAI | 1 433 | 1 426–1 440 | 10 459 | Proprietary |
| 54 | Kimi K2.6 | Moonshot AI | 1 432 | 1 423–1 440 | 5 821 | Modified MIT |
| 55 | GPT-5.4 | OpenAI | 1 429 | 1 422–1 436 | 10 305 | Proprietary |
| 56 | GPT-6 Astra в режиме max | OpenAI | 1 429 | 1 402–1 455 | 572 | Proprietary |
| 57 | GPT-5.1 в режиме high | OpenAI | 1 427 | 1 419–1 435 | 5 967 | Proprietary |
| 58 | ERNIE 5.0 · 0110 | Baidu | 1 426 | 1 417–1 434 | 5 454 | Proprietary |
| 59 | Claude Sonnet 4.5 в режиме high | Anthropic | 1 424 | 1 418–1 430 | 12 117 | Proprietary |
| 60 | ERNIE 5.0 · preview-1203 | Baidu | 1 424 | 1 408–1 439 | 1 523 | Proprietary |
| 61 | Kimi K2.5 · thinking | Moonshot AI | 1 422 | 1 416–1 429 | 11 367 | Modified MIT |
| 62 | GPT-5.5 · 5.5-instant | OpenAI | 1 419 | 1 409–1 429 | 4 111 | Proprietary |
| 63 | Gemini 3.5 Flash Lite | 1 419 | 1 410–1 428 | 5 305 | Proprietary | |
| 64 | Claude Sonnet 5 в режиме high | Anthropic | 1 417 | 1 409–1 426 | 6 914 | Proprietary |
| 65 | MiMo-V2 Pro | Xiaomi | 1 417 | 1 407–1 427 | 3 565 | Proprietary |
| 66 | Gemma 4 31B | 1 416 | 1 397–1 435 | 930 | Apache 2.0 | |
| 67 | ERNIE 5.0 · preview-1022 | Baidu | 1 416 | 1 394–1 438 | 719 | Proprietary |
| 68 | Grok 3 | xAI | 1 414 | 1 405–1 423 | 4 697 | Proprietary |
| 69 | GLM 4.6 | Z.ai | 1 412 | 1 403–1 420 | 5 013 | MIT |
| 70 | Claude Opus 4.1 · 20250805-thinking-16k | Anthropic | 1 411 | 1 403–1 419 | 6 734 | Proprietary |
| 71 | Grok 4.1 · 4.1 | xAI | 1 411 | 1 404–1 418 | 9 641 | Proprietary |
| 72 | GLM 5V Turbo | Z.ai | 1 410 | 1 395–1 426 | 1 653 | Proprietary |
| 73 | Claude Opus 4.1 · 20250805 | Anthropic | 1 410 | 1 404–1 416 | 10 589 | Proprietary |
| 74 | GPT-5.6 Terra в режиме xhigh | OpenAI | 1 409 | 1 400–1 418 | 5 723 | Proprietary |
| 75 | ChatGPT-4o (latest) | OpenAI | 1 406 | 1 400–1 412 | 11 308 | Proprietary |
| 76 | R1 0528 | DeepSeek | 1 406 | 1 393–1 418 | 2 391 | MIT |
| 77 | Gemma 4 26B A4B | 1 405 | 1 387–1 424 | 950 | Apache 2.0 | |
| 78 | Grok 4.1 · 4.1-thinking | xAI | 1 405 | 1 399–1 412 | 9 705 | Proprietary |
| 79 | MiniMax M3 | MiniMax | 1 405 | 1 398–1 413 | 8 529 | MiniMax Community License |
| 80 | Seed 2.0 Pro | ByteDance Seed | 1 405 | 1 399–1 412 | 12 122 | Proprietary |
| 81 | Qwen3.6 Plus | Alibaba Qwen | 1 404 | 1 396–1 412 | 6 671 | Proprietary |
| 82 | GPT-5.1 | OpenAI | 1 404 | 1 396–1 412 | 6 397 | Proprietary |
| 83 | DeepSeek V4 Flash 0423 | DeepSeek | 1 403 | 1 395–1 411 | 7 953 | MIT |
| 84 | DeepSeek V3.2 Exp · exp | DeepSeek | 1 403 | 1 388–1 417 | 1 618 | MIT |
| 85 | DeepSeek V3.1 Terminus · terminus | DeepSeek | 1 402 | 1 375–1 429 | 450 | MIT |
| 86 | Qwen3.5 397B A17B | Alibaba Qwen | 1 402 | 1 396–1 408 | 12 803 | Apache 2.0 |
| 87 | GPT-5.2 Chat | OpenAI | 1 402 | 1 393–1 410 | 5 314 | Proprietary |
| 88 | Gemini 2.5 Flash · flash | 1 402 | 1 397–1 407 | 17 148 | Proprietary | |
| 89 | DeepSeek V3.1 · v3.1-thinking | DeepSeek | 1 401 | 1 387–1 416 | 1 616 | MIT |
| 90 | Gemini 3.1 Flash Lite Preview | 1 400 | 1 393–1 408 | 9 740 | Proprietary | |
| 91 | GLM 4.7 | Z.ai | 1 400 | 1 387–1 413 | 1 890 | MIT |
| 92 | Qwen3 Max · preview | Alibaba Qwen | 1 400 | 1 390–1 410 | 3 685 | Proprietary |
| 93 | DeepSeek V3.2 · v3.2 | DeepSeek | 1 400 | 1 392–1 408 | 6 610 | MIT |
| 94 | Hy3 | Tencent | 1 399 | 1 383–1 414 | 1 533 | Apache 2.0 |
| 95 | DeepSeek V4 Flash 0423 в режиме high | DeepSeek | 1 397 | 1 390–1 405 | 7 985 | MIT |
| 96 | Grok 4 | xAI | 1 397 | 1 388–1 405 | 5 258 | Proprietary |
| 97 | GPT-5.6 Luna в режиме xhigh | OpenAI | 1 396 | 1 387–1 405 | 5 652 | Proprietary |
| 98 | GLM 4.5 | Z.ai | 1 395 | 1 384–1 406 | 3 075 | MIT |
| 99 | DeepSeek V3.1 Terminus · terminus-thinking | DeepSeek | 1 395 | 1 367–1 423 | 432 | MIT |
| 100 | Grok 4.1 Fast | xAI | 1 394 | 1 387–1 401 | 8 599 | Proprietary |
| 101 | GPT-4.5 Preview | OpenAI | 1 394 | 1 382–1 406 | 2 618 | Proprietary |
| 102 | Hunyuan T1 | Tencent | 1 393 | 1 369–1 416 | 614 | Proprietary |
| 103 | DeepSeek V3.2 Exp · exp-thinking | DeepSeek | 1 392 | 1 375–1 410 | 1 188 | MIT |
| 104 | Nemotron 3 Ultra | NVIDIA | 1 392 | 1 378–1 406 | 1 917 | OpenMDW-1.1 |
| 105 | MiMo-V2 Omni | Xiaomi | 1 392 | 1 381–1 403 | 3 213 | Proprietary |
| 106 | Mistral Medium 3.1 | Mistral AI | 1 392 | 1 386–1 397 | 13 574 | Proprietary |
| 107 | MiMo-V2.5 | Xiaomi | 1 391 | 1 383–1 399 | 7 315 | MIT |
| 108 | DeepSeek V3.1 · v3.1 | DeepSeek | 1 389 | 1 375–1 402 | 1 977 | MIT |
| 109 | DeepSeek V3.2 · v3.2-thinking | DeepSeek | 1 389 | 1 381–1 397 | 6 135 | MIT |
| 110 | LongCat-Flash-Chat · chat-2602-exp | Meituan | 1 389 | 1 379–1 398 | 4 203 | Proprietary |
| 111 | Grok 4 Fast · chat | xAI | 1 388 | 1 368–1 408 | 886 | Proprietary |
| 112 | Gemini 2.5 Flash · flash-preview-09-2025 | 1 387 | 1 378–1 396 | 4 436 | Proprietary | |
| 113 | Inkling | Thinking Machines Lab | 1 387 | 1 378–1 396 | 5 269 | Apache 2.0 |
| 114 | Mistral Large 3 2512 | Mistral AI | 1 386 | 1 380–1 393 | 11 713 | Apache 2.0 |
| 115 | Hunyuan Vision 1.5 | Tencent | 1 386 | 1 350–1 422 | 254 | Proprietary |
| 116 | Claude Opus 4 · 20250514-thinking-16k | Anthropic | 1 386 | 1 377–1 395 | 4 434 | Proprietary |
| 117 | Qwen3 235B A22B Thinking 2507 | Alibaba Qwen | 1 385 | 1 367–1 402 | 1 067 | Apache 2.0 |
| 118 | Grok 4.3 | xAI | 1 382 | 1 376–1 389 | 11 969 | Proprietary |
| 119 | Qwen3 Max · 2025-09-23 | Alibaba Qwen | 1 381 | 1 364–1 398 | 1 133 | Proprietary |
| 120 | Kimi K2.5 · instant | Moonshot AI | 1 380 | 1 363–1 396 | 1 244 | Modified MIT |
| 121 | Qwen3.8 27B | Alibaba Qwen | 1 378 | 1 365–1 392 | 2 059 | Apache 2.0 |
| 122 | GPT-5.2 в режиме high | OpenAI | 1 377 | 1 370–1 385 | 7 278 | Proprietary |
| 123 | MiMo-V2 Flash · flash (non-thinking) | Xiaomi | 1 377 | 1 370–1 385 | 7 008 | MIT |
| 124 | GPT-5.2 | OpenAI | 1 376 | 1 370–1 382 | 12 860 | Proprietary |
| 125 | Qwen3 235B A22B Instruct 2507 | Alibaba Qwen | 1 375 | 1 370–1 381 | 13 441 | Apache 2.0 |
| 126 | Grok 4 Fast · reasoning | xAI | 1 375 | 1 363–1 387 | 2 447 | Proprietary |
| 127 | Claude Opus 4 · 20250514 | Anthropic | 1 374 | 1 366–1 383 | 5 371 | Proprietary |
| 128 | Kimi K2 Thinking | Moonshot AI | 1 374 | 1 367–1 381 | 9 070 | Modified MIT |
| 129 | Claude Haiku 4.5 | Anthropic | 1 372 | 1 367–1 377 | 20 900 | Proprietary |
| 130 | GPT-5.4 Mini в режиме high | OpenAI | 1 372 | 1 364–1 379 | 9 804 | Proprietary |
| 131 | Mistral Medium 3.5 | Mistral AI | 1 370 | 1 356–1 385 | 1 860 | Modified MIT |
| 132 | Qwen3.5-122B-A10B | Alibaba Qwen | 1 369 | 1 359–1 378 | 4 295 | Apache 2.0 |
| 133 | Amazon Nova Experimental Chat · 26-02-10 | Amazon | 1 366 | 1 340–1 393 | 455 | Proprietary |
| 134 | Qwen3 VL 235B A22B Instruct | Alibaba Qwen | 1 366 | 1 349–1 383 | 1 342 | Apache 2.0 |
| 135 | GPT-5 | OpenAI | 1 366 | 1 357–1 376 | 3 962 | Proprietary |
| 136 | Gemini 2.5 Flash Lite · 06-17-thinking | 1 366 | 1 357–1 375 | 4 126 | Proprietary | |
| 137 | GPT-5 в режиме high | OpenAI | 1 365 | 1 355–1 374 | 4 192 | Proprietary |
| 138 | DeepSeek V3 0324 | DeepSeek | 1 364 | 1 356–1 372 | 5 970 | MIT |
| 139 | GPT-4.1 | OpenAI | 1 364 | 1 356–1 371 | 6 595 | Proprietary |
| 140 | MiniMax M2.1 | MiniMax | 1 362 | 1 351–1 374 | 2 679 | MIT |
| 141 | Qwen3.5-27B | Alibaba Qwen | 1 361 | 1 351–1 370 | 4 050 | Apache 2.0 |
| 142 | Gemini 2.5 Flash Lite · 09-2025-no-thinking | 1 359 | 1 351–1 367 | 6 417 | Proprietary | |
| 143 | o3 | OpenAI | 1 358 | 1 351–1 366 | 7 576 | Proprietary |
| 144 | Hunyuan TurboS · 20250416 | Tencent | 1 358 | 1 343–1 374 | 1 474 | Proprietary |
| 145 | Step 3.5 Flash | StepFun | 1 357 | 1 350–1 363 | 9 159 | Apache 2.0 |
| 146 | Qwen3 235B A22B · a22b-no-thinking | Alibaba Qwen | 1 356 | 1 347–1 365 | 4 546 | Apache 2.0 |
| 147 | GPT-5.3 Chat | OpenAI | 1 355 | 1 346–1 364 | 5 082 | Proprietary |
| 148 | R1 | DeepSeek | 1 354 | 1 344–1 364 | 3 289 | MIT |
| 149 | MiniMax M2.7 | MiniMax | 1 351 | 1 345–1 358 | 11 370 | Modified MIT |
| 150 | Amazon Nova Experimental Chat · 12-10 | Amazon | 1 351 | 1 327–1 375 | 576 | Proprietary |
| 151 | Amazon Nova Experimental Chat · 11-10 | Amazon | 1 349 | 1 340–1 359 | 3 798 | Proprietary |
| 152 | MiMo-V2 Flash · flash (thinking) | Xiaomi | 1 348 | 1 334–1 363 | 1 668 | MIT |
| 153 | Qwen3.5-35B-A3B | Alibaba Qwen | 1 348 | 1 339–1 357 | 4 432 | Apache 2.0 |
| 154 | Muse Glimmer 30B | Meta | 1 347 | 1 324–1 371 | 693 | Apache-2.0 |
| 155 | o1 · 2024-12-17 | OpenAI | 1 347 | 1 338–1 356 | 4 642 | Proprietary |
| 156 | Kimi K2 0905 | Moonshot AI | 1 347 | 1 332–1 362 | 1 482 | Modified MIT |
| 157 | LongCat-Flash-Chat · chat | Meituan | 1 347 | 1 331–1 363 | 1 350 | MIT |
| 158 | GLM 4.6V | Z.ai | 1 346 | 1 317–1 375 | 406 | MIT |
| 159 | Gemma 3 27B | 1 345 | 1 338–1 353 | 6 272 | Gemma | |
| 160 | Qwen3 VL 235B A22B Thinking | Alibaba Qwen | 1 345 | 1 327–1 364 | 1 009 | Apache 2.0 |
| 161 | Claude Sonnet 4 · 20250514-thinking-32k | Anthropic | 1 345 | 1 336–1 355 | 4 227 | Proprietary |
| 162 | Mistral Medium 3 | Mistral AI | 1 343 | 1 334–1 353 | 4 043 | Proprietary |
| 163 | Qwen3.5-Flash | Alibaba Qwen | 1 343 | 1 336–1 350 | 9 033 | Proprietary |
| 164 | GLM 4.5 Air | Z.ai | 1 342 | 1 333–1 352 | 3 938 | MIT |
| 165 | Grok 3 Mini | xAI | 1 342 | 1 331–1 354 | 2 546 | Proprietary |
| 166 | Claude Sonnet 4 · 20250514 | Anthropic | 1 342 | 1 333–1 351 | 4 879 | Proprietary |
| 167 | Hy3 preview | Tencent | 1 341 | 1 322–1 361 | 999 | tencent-hunyuan-community |
| 168 | Gemini 2.0 Flash | 1 340 | 1 332–1 347 | 6 257 | Proprietary | |
| 169 | Qwen2.5 Max | Alibaba Qwen | 1 339 | 1 330–1 347 | 4 994 | Proprietary |
| 170 | Grok 3 Mini в режиме high | xAI | 1 334 | 1 320–1 347 | 1 893 | Proprietary |
| 171 | Qwen3 Coder 480B A35B | Alibaba Qwen | 1 333 | 1 323–1 344 | 3 213 | Apache 2.0 |
| 172 | Amazon Nova Experimental Chat · 26-01-10 | Amazon | 1 333 | 1 308–1 358 | 545 | Proprietary |
| 173 | Qwen3 Next 80B A3B Instruct | Alibaba Qwen | 1 333 | 1 322–1 343 | 2 990 | Apache 2.0 |
| 174 | Gemini 1.5 Pro · 1.5-pro-002 | 1 333 | 1 325–1 340 | 8 062 | Proprietary | |
| 175 | MiniMax M2.5 | MiniMax | 1 332 | 1 324–1 340 | 6 319 | Modified MIT |
| 176 | Claude 3.7 Sonnet · 20250219-thinking-32k | Anthropic | 1 332 | 1 323–1 341 | 5 127 | Proprietary |
| 177 | Gemma 3 12B | 1 331 | 1 308–1 354 | 600 | Gemma | |
| 178 | DeepSeek V3 | DeepSeek | 1 329 | 1 319–1 339 | 3 623 | DeepSeek |
| 179 | Step-2 16k | StepFun | 1 327 | 1 307–1 348 | 751 | Proprietary |
| 180 | Inkling Small | Thinking Machines Lab | 1 325 | 1 315–1 336 | 3 859 | Apache 2.0 |
| 181 | GPT-5 Mini в режиме high | OpenAI | 1 325 | 1 314–1 335 | 3 494 | Proprietary |
| 182 | Kimi K2 0711 | Moonshot AI | 1 324 | 1 313–1 334 | 3 383 | Modified MIT |
| 183 | Step 3 | StepFun | 1 321 | 1 300–1 342 | 766 | Apache 2.0 |
| 184 | Trinity Large | Arcee AI | 1 320 | 1 311–1 330 | 4 527 | Apache 2.0 |
| 185 | INTELLECT-3 | Prime Intellect | 1 319 | 1 298–1 340 | 795 | MIT |
| 186 | Gemini 2.0 Flash-Lite | 1 319 | 1 310–1 328 | 4 104 | Proprietary | |
| 187 | Command A | Cohere | 1 319 | 1 311–1 326 | 7 549 | CC-BY-NC-4.0 |
| 188 | Qwen3 30B A3B Instruct 2507 | Alibaba Qwen | 1 318 | 1 308–1 329 | 2 947 | Apache 2.0 |
| 189 | o1 · preview | OpenAI | 1 318 | 1 308–1 328 | 4 508 | Proprietary |
| 190 | Amazon Nova Experimental Chat · 10-20 | Amazon | 1 317 | 1 302–1 333 | 1 534 | Proprietary |
| 191 | Claude 3.7 Sonnet · 20250219 | Anthropic | 1 317 | 1 309–1 325 | 5 905 | Proprietary |
| 192 | Qwen3 235B A22B · a22b | Alibaba Qwen | 1 317 | 1 306–1 327 | 3 544 | Apache 2.0 |
| 193 | GPT-5.4 Nano в режиме high | OpenAI | 1 316 | 1 309–1 323 | 9 445 | Proprietary |
| 194 | Nemotron 3 Super | NVIDIA | 1 315 | 1 297–1 334 | 1 057 | NVIDIA Open Model |
| 195 | Qwen3 Next 80B A3B Thinking | Alibaba Qwen | 1 314 | 1 300–1 329 | 1 755 | Apache 2.0 |
| 196 | Llama 3.1 Nemotron Ultra 253B v1 | NVIDIA | 1 314 | 1 288–1 341 | 444 | Nvidia Open Model |
| 197 | GLM-4-Plus · plus-0111 | Z.ai | 1 311 | 1 292–1 330 | 918 | Proprietary |
| 198 | Trinity Large Thinking | Arcee AI | 1 308 | 1 299–1 318 | 4 498 | Apache 2.0 |
| 199 | Step-1o Turbo | StepFun | 1 308 | 1 289–1 327 | 981 | Proprietary |
| 200 | Llama 3.3 Nemotron Super 49B v1.5 | NVIDIA | 1 308 | 1 279–1 337 | 366 | Nvidia Open |
| 201 | Amazon Nova Experimental Chat · 10-09 | Amazon | 1 306 | 1 275–1 338 | 322 | Proprietary |
| 202 | Mistral Small 3.2 24B | Mistral AI | 1 304 | 1 291–1 317 | 2 131 | Apache 2.0 |
| 203 | GPT-4.1 Mini | OpenAI | 1 300 | 1 291–1 308 | 5 079 | Proprietary |
| 204 | MiniMax M1 | MiniMax | 1 298 | 1 289–1 307 | 4 448 | Apache 2.0 |
| 205 | Qwen3 32B | Alibaba Qwen | 1 297 | 1 275–1 320 | 616 | Apache 2.0 |
| 206 | GLM 4.7 Flash | Z.ai | 1 297 | 1 283–1 311 | 1 772 | MIT |
| 207 | Mercury 2 | Inception Labs | 1 295 | 1 269–1 321 | 514 | Proprietary |
| 208 | o4 Mini | OpenAI | 1 294 | 1 286–1 302 | 5 747 | Proprietary |
| 209 | GLM 4.5V | Z.ai | 1 294 | 1 271–1 317 | 613 | MIT |
| 210 | Nova 2 Lite | Amazon | 1 293 | 1 278–1 307 | 1 713 | Proprietary |
| 211 | Qwen-Plus | Alibaba Qwen | 1 292 | 1 274–1 311 | 931 | Proprietary |
| 212 | Claude 3.5 Sonnet · 20241022 | Anthropic | 1 292 | 1 286–1 298 | 11 696 | Proprietary |
| 213 | GPT-4o (2024-05-13) | OpenAI | 1 291 | 1 284–1 298 | 16 599 | Proprietary |
| 214 | Granite 4.2 30B | IBM Granite | 1 290 | 1 264–1 316 | 604 | Apache 2.0 |
| 215 | MiniMax M2 | MiniMax | 1 289 | 1 269–1 309 | 887 | Apache 2.0 |
| 216 | Solar Pro 4 | Upstage | 1 288 | 1 261–1 316 | 523 | Proprietary |
| 217 | Llama 3.3 Nemotron Super 49B v1 | NVIDIA | 1 288 | 1 261–1 315 | 428 | Nvidia |
| 218 | QwQ 32B | Alibaba Qwen | 1 287 | 1 278–1 297 | 3 658 | Apache 2.0 |
| 219 | Gemma 3n E4B | 1 287 | 1 276–1 298 | 2 847 | Gemma | |
| 220 | o3 Mini High | OpenAI | 1 286 | 1 275–1 296 | 3 065 | Proprietary |
| 221 | Ling-flash-2.0 | inclusionAI | 1 285 | 1 265–1 306 | 832 | MIT |
| 222 | Gemini 1.5 Flash · 002 | 1 285 | 1 276–1 294 | 5 017 | Proprietary | |
| 223 | DeepSeek V2.5 · v2.5-1210 | DeepSeek | 1 285 | 1 268–1 302 | 1 090 | DeepSeek |
| 224 | Grok 2 | xAI | 1 284 | 1 277–1 292 | 9 339 | Proprietary |
| 225 | Gemini 1.5 Pro · advanced-0514 | 1 284 | 1 274–1 294 | 7 025 | Proprietary | |
| 226 | Yi-Lightning | 01.AI | 1 280 | 1 270–1 290 | 3 781 | Proprietary |
| 227 | GPT-4o (2024-08-06) | OpenAI | 1 274 | 1 266–1 283 | 6 756 | Proprietary |
| 228 | gpt-oss-120b | OpenAI | 1 274 | 1 264–1 283 | 3 872 | Apache 2.0 |
| 229 | o3 Mini | OpenAI | 1 273 | 1 266–1 280 | 8 138 | Proprietary |
| 230 | Gemini 1.5 Pro · 1.5-pro-001 | 1 272 | 1 264–1 280 | 11 797 | Proprietary | |
| 231 | Nemotron 3 Nano 30B A3B | NVIDIA | 1 272 | 1 259–1 285 | 2 315 | NVIDIA Open Model |
| 232 | Ring-flash-2.0 | inclusionAI | 1 272 | 1 252–1 292 | 860 | MIT |
| 233 | Qwen3 30B A3B | Alibaba Qwen | 1 271 | 1 261–1 281 | 3 328 | Apache 2.0 |
| 234 | Gemma 3 4B | 1 270 | 1 249–1 292 | 666 | Gemma | |
| 235 | Hunyuan TurboS · 20250226 | Tencent | 1 270 | 1 245–1 296 | 424 | Proprietary |
| 236 | Hunyuan Turbo | Tencent | 1 269 | 1 244–1 294 | 451 | Proprietary |
| 237 | Llama 3.1 Nemotron 70B Instruct | NVIDIA | 1 268 | 1 250–1 287 | 1 001 | Llama 3.1 |
| 238 | GPT-4 Turbo | OpenAI | 1 268 | 1 261–1 276 | 14 396 | Proprietary |
| 239 | GPT-4o-mini (2024-07-18) | OpenAI | 1 268 | 1 261–1 275 | 10 484 | Proprietary |
| 240 | Llama 4 Maverick | Meta | 1 266 | 1 258–1 275 | 5 129 | Llama 4 |
| 241 | OLMo 3.1 32B Instruct | Ai2 | 1 265 | 1 250–1 280 | 1 637 | Apache 2.0 |
| 242 | Hunyuan Large | Tencent | 1 264 | 1 241–1 288 | 568 | Proprietary |
| 243 | GLM-4-Plus · plus | Z.ai | 1 263 | 1 253–1 274 | 3 801 | Proprietary |
| 244 | Llama 3.1 405B Instruct · fp8 | Meta | 1 262 | 1 254–1 270 | 8 948 | Llama 3.1 Community |
| 245 | Qwen2.5 Plus | Alibaba Qwen | 1 262 | 1 247–1 276 | 1 591 | Proprietary |
| 246 | GPT-4.1 Nano | OpenAI | 1 260 | 1 242–1 278 | 1 005 | Proprietary |
| 247 | Llama 3.1 405B Instruct · bf16 | Meta | 1 260 | 1 252–1 267 | 6 512 | Llama 3.1 Community |
| 248 | Granite 4.1 8B | IBM Granite | 1 255 | 1 229–1 280 | 664 | Apache 2.0 |
| 249 | OLMo 3 32B Think | Ai2 | 1 255 | 1 234–1 275 | 849 | Apache 2.0 |
| 250 | Nemotron 3.5 Lightning | NVIDIA | 1 252 | 1 236–1 268 | 1 750 | OpenMDW-1.1 |
| 251 | Mistral Small 3.1 24B | Mistral AI | 1 251 | 1 242–1 261 | 4 101 | Apache 2.0 |
| 252 | Llama 3.3 70B Instruct | Meta | 1 250 | 1 243–1 257 | 8 024 | Llama-3.3 |
| 253 | Hunyuan Large Vision | Tencent | 1 250 | 1 226–1 274 | 623 | Proprietary |
| 254 | GPT-5 Nano в режиме high | OpenAI | 1 250 | 1 230–1 269 | 911 | Proprietary |
| 255 | Llama 4 Scout | Meta | 1 249 | 1 239–1 259 | 3 743 | Llama |
| 256 | Qwen Max | Alibaba Qwen | 1 248 | 1 236–1 260 | 2 358 | Qwen |
| 257 | Magistral Medium | Mistral AI | 1 245 | 1 229–1 262 | 1 342 | Proprietary |
| 258 | GPT-4 Turbo Preview · 1106-preview | OpenAI | 1 244 | 1 236–1 252 | 15 576 | Proprietary |
| 259 | o1-mini | OpenAI | 1 244 | 1 237–1 252 | 7 858 | Proprietary |
| 260 | Mistral Large 2407 | Mistral AI | 1 243 | 1 234–1 251 | 6 692 | Mistral Research |
| 261 | Hunyuan Standard | Tencent | 1 242 | 1 219–1 265 | 590 | Proprietary |
| 262 | Mistral Large | Mistral AI | 1 242 | 1 233–1 251 | 4 439 | MRL |
| 263 | Grok 2 Mini | xAI | 1 241 | 1 234–1 249 | 7 924 | Proprietary |
| 264 | Gemma 2 27B | 1 241 | 1 234–1 247 | 11 429 | Gemma license | |
| 265 | Granite 4.2 8B | IBM Granite | 1 240 | 1 212–1 268 | 585 | Apache 2.0 |
| 266 | Gemma 2 9B IT SimPO | Princeton NLP | 1 239 | 1 224–1 253 | 1 671 | MIT |
| 267 | Athene 70B | Nexusflow | 1 238 | 1 227–1 249 | 3 097 | CC-BY-NC-4.0 |
| 268 | Claude 3.5 Sonnet · 20240620 | Anthropic | 1 238 | 1 230–1 246 | 12 430 | Proprietary |
| 269 | DeepSeek V2.5 · v2.5 | DeepSeek | 1 237 | 1 226–1 247 | 3 450 | DeepSeek |
| 270 | Reka Core | Reka AI | 1 236 | 1 218–1 254 | 1 037 | Proprietary |
| 271 | Claude 3 Opus | Anthropic | 1 235 | 1 228–1 241 | 28 395 | Proprietary |
| 272 | Command R+ (08-2024) | Cohere | 1 235 | 1 220–1 250 | 1 474 | CC-BY-NC-4.0 |
| 273 | Athene V2 Chat | Nexusflow | 1 234 | 1 224–1 244 | 3 713 | NexusFlow |
| 274 | Claude 3.5 Haiku | Anthropic | 1 232 | 1 226–1 239 | 9 815 | Proprietary |
| 275 | GPT-4 Turbo Preview · 0125-preview | OpenAI | 1 232 | 1 224–1 240 | 13 940 | Proprietary |
| 276 | Llama 3.1 70B Instruct | Meta | 1 232 | 1 224–1 240 | 8 250 | Llama 3.1 Community |
| 277 | Llama 3.1 Tulu 3 70B | Ai2 | 1 231 | 1 207–1 256 | 441 | Llama 3.1 |
| 278 | OLMo 3.1 32B Think | Ai2 | 1 225 | 1 207–1 242 | 1 322 | Apache 2.0 |
| 279 | Gemini 1.5 Flash · 001 | 1 222 | 1 214–1 230 | 9 031 | Proprietary | |
| 280 | Qwen2.5 72B Instruct | Alibaba Qwen | 1 221 | 1 213–1 230 | 5 684 | Qwen |
| 281 | Granite 4.2 3B | IBM Granite | 1 220 | 1 191–1 249 | 562 | Apache 2.0 |
| 282 | Gemini 1.5 Flash-8B | 1 217 | 1 209–1 226 | 5 031 | Proprietary | |
| 283 | Llama 3.1 Nemotron 51B Instruct | NVIDIA | 1 213 | 1 188–1 239 | 482 | Llama 3.1 |
| 284 | Nova Pro 1.0 | Amazon | 1 212 | 1 203–1 222 | 3 944 | Proprietary |
| 285 | Jamba 1.5 Large | AI21 Labs | 1 212 | 1 196–1 228 | 1 379 | Jamba Open |
| 286 | Llama 3 70B Instruct | Meta | 1 210 | 1 202–1 217 | 22 753 | Llama 3 Community |
| 287 | Granite 4.0 H Small | IBM Granite | 1 209 | 1 186–1 233 | 682 | Apache 2.0 |
| 288 | Gemma 2 9B | 1 205 | 1 198–1 213 | 8 340 | Gemma license | |
| 289 | Nemotron-4 340B Instruct | NVIDIA | 1 202 | 1 191–1 214 | 3 152 | NVIDIA Open Model |
| 290 | Reka Flash (2024-09) | Reka AI | 1 202 | 1 185–1 220 | 1 037 | Proprietary |
| 291 | gpt-oss-20b | OpenAI | 1 201 | 1 183–1 219 | 1 227 | Apache 2.0 |
| 292 | Command R+ | Cohere | 1 201 | 1 192–1 209 | 11 012 | CC-BY-NC-4.0 |
| 293 | Aya Expanse 32B | Cohere | 1 200 | 1 190–1 210 | 3 853 | CC-BY-NC-4.0 |
| 294 | OLMo 2 32B Instruct | Ai2 | 1 198 | 1 174–1 223 | 567 | Apache-2.0 |
| 295 | GLM-4 | Z.ai | 1 198 | 1 183–1 213 | 1 547 | Proprietary |
| 296 | Nova Lite 1.0 | Amazon | 1 197 | 1 186–1 207 | 3 037 | Proprietary |
| 297 | Mistral Small 3 | Mistral AI | 1 195 | 1 182–1 207 | 2 415 | Apache 2.0 |
| 298 | GPT-4 · 0613 | OpenAI | 1 191 | 1 183–1 200 | 13 932 | Proprietary |
| 299 | Mercury | Inception Labs | 1 190 | 1 152–1 228 | 293 | Proprietary |
| 300 | GPT-4 · 0314 | OpenAI | 1 190 | 1 180–1 200 | 8 231 | Proprietary |
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 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.