Сводный рейтинг и рейтинги по задачам. Счёт по шкале BenchLM (0–100), «подтверждено» значит оценка опирается на опубликованные бенчмарки.
| # | Модель | Разработчик | Тип | Реж. | Счёт | Статус |
|---|---|---|---|---|---|---|
| #1 | Gemini 4 Carbon | api | reasoning | 88.38 | ● оценка | |
| #2 | Claude Opus 5.5 | Anthropic | api | reasoning | 86.29 | ● подтверждено |
| #3 | GPT-6 Astra | OpenAI | api | reasoning | 86.07 | ● подтверждено |
| #4 | GPT-6.1 Sol | OpenAI | api | reasoning | 85.27 | ● подтверждено |
| #5 | Gemini 4 Argon | api | reasoning | 82.28 | ● подтверждено | |
| #6 | Claude Sonnet 5.5 | Anthropic | api | reasoning | 81.84 | ● подтверждено |
| #7 | Claude Fable 5.1 | Anthropic | api | reasoning | 81.08 | ● подтверждено |
| #8 | Claude Fable 5 | Anthropic | api | reasoning | 80.78 | ● подтверждено |
| #9 | GPT-6.1 Terra | OpenAI | api | reasoning | 80.46 | ● оценка |
| #10 | Claude Opus 5 | Anthropic | api | reasoning | 79.68 | ● подтверждено |
| #11 | GPT-6 Sol | OpenAI | api | reasoning | 77.97 | ● подтверждено |
| #12 | GPT-5.6 Sol | OpenAI | api | reasoning | 74.83 | ● подтверждено |
| #13 | Kimi K3.1 Pro | Moonshot AI | — | reasoning | 74.61 | ● оценка |
| #14 | GPT-5.5 Pro | OpenAI | api | reasoning | 72.76 | ● оценка |
| #15 | GPT-5.6 Terra | OpenAI | api | reasoning | 69.81 | ● подтверждено |
| #16 | Claude Opus 4.8 | Anthropic | api | reasoning | 69.63 | ● подтверждено |
| #17 | Kimi K3 | Moonshot AI | open | reasoning | 69.34 | ● подтверждено |
| #18 | Muse Wonder Full | Meta | api | reasoning | 68.96 | ● оценка |
| #19 | GPT-5.5 | OpenAI | api | reasoning | 68.82 | ● подтверждено |
| #20 | GPT-5.4 Pro | OpenAI | api | reasoning | 68.23 | ● оценка |
| #21 | Claude Opus 4.7 (Adaptive) | Anthropic | api | reasoning | 68.11 | ● оценка |
| #22 | Grok 4.6 | xAI | api | reasoning | 66.62 | ● подтверждено |
| #23 | Gemini 3.7 Flash | api | reasoning | 66.53 | ● подтверждено | |
| #24 | Claude Haiku 5.5 | Anthropic | api | reasoning | 66.42 | ● подтверждено |
| #25 | GPT-5.4 | OpenAI | api | reasoning | 66.4 | ● подтверждено |
| #26 | GPT-5.2 Pro | OpenAI | api | reasoning | 66.23 | ● оценка |
| #27 | Gemini 3.8 Flash | api | reasoning | 66.22 | ● подтверждено | |
| #28 | Muse Spark 1.2 | Meta | api | reasoning | 65.41 | ● подтверждено |
| #29 | Ling 3.1 Flash | InclusionAI | open | reasoning | 65.41 | ● оценка |
| #30 | DeepSeek V4.1 Flash 1017 | DeepSeek | open | reasoning | 65.22 | ● подтверждено |
| #31 | Claude Opus 4.7 | Anthropic | api | 65.18 | ● оценка | |
| #32 | Muse Spark 1.1 | Meta | api | reasoning | 65.16 | ● подтверждено |
| #33 | Qwen3.8 Max | Alibaba | open | reasoning | 64.86 | ● подтверждено |
| #34 | Gemini 3.1 Pro | api | reasoning | 64.41 | ● оценка | |
| #35 | dots3-note Preview | Dots Studio | open | reasoning | 64.35 | ● оценка |
| #36 | Claude Sonnet 5 | Anthropic | api | reasoning | 64.25 | ● подтверждено |
| #37 | GPT-5.6 Luna | OpenAI | api | reasoning | 63.8 | ● подтверждено |
| #38 | Sage-A1.5 | InternScience | open | reasoning | 62.98 | ● оценка |
| #39 | Grok 4.5 | xAI | api | reasoning | 62.92 | ● подтверждено |
| #40 | Muse Wonder | Meta | api | reasoning | 62.78 | ● оценка |
| #41 | GPT-5.3 Codex | OpenAI | api | reasoning | 62.75 | ● оценка |
| #42 | DeepSeek V4 Pro 0813 | DeepSeek | open | reasoning | 62.53 | ● оценка |
| #43 | Ornith-1.5-397B | Ornith AI | open | reasoning | 62.47 | ● оценка |
| #44 | Gemini 3.6 Flash | api | reasoning | 62.41 | ● подтверждено | |
| #45 | Qwen3.7 Max | Alibaba | api | reasoning | 62.32 | ● подтверждено |
| #46 | GPT-6 Luna | OpenAI | api | reasoning | 62.14 | ● подтверждено |
| #47 | Claude Opus 4.6 | Anthropic | api | 61.82 | ● подтверждено | |
| #48 | Hy4 preview | Tencent | open | reasoning | 61.46 | ● оценка |
| #49 | Gemini 3.5 Flash | api | reasoning | 61.4 | ● подтверждено | |
| #50 | GLM-5.3 | Z.AI | open | reasoning | 61.03 | ● подтверждено |
| #51 | Gemini 3 Pro | api | 60.68 | ● подтверждено | ||
| #52 | GLM-5.2 | Z.AI | open | reasoning | 60.03 | ● подтверждено |
| #53 | Muse Spark | Meta | api | reasoning | 59.84 | ● подтверждено |
| #54 | MiniMax M3.1 Flash | MiniMax | api | reasoning | 59.71 | ● оценка |
| #55 | GPT-5.2 | OpenAI | api | reasoning | 59.41 | ● подтверждено |
| #56 | GLM-5.3-Flash | Z.AI | open | reasoning | 58.26 | ● оценка |
| #57 | Kimi K2.6 | Moonshot AI | open | reasoning | 57.38 | ● оценка |
| #58 | GPT-5.1 | OpenAI | api | reasoning | 56.8 | ● подтверждено |
| #59 | MiMo-V2.6-Pro | Xiaomi | open | reasoning | 56.44 | ● оценка |
| #60 | GPT-5.2-Codex | OpenAI | api | reasoning | 56.25 | ● оценка |
| #61 | Kimi K2.5 (Reasoning) | Moonshot AI | api | reasoning | 56.24 | ● оценка |
| #62 | Step 5 Preview | StepFun | open | reasoning | 56.15 | ● оценка |
| #63 | Grok 4.20 | xAI | api | reasoning | 55.99 | ● оценка |
| #64 | GLM-5.1 | Z.AI | open | reasoning | 55.83 | ● подтверждено |
| #65 | Qwen3.8-Flash-Next | Alibaba | open | reasoning | 55.8 | ● оценка |
| #66 | Agents-A1 | InternScience | open | reasoning | 55.74 | ● оценка |
| #67 | Agents-A1-F16-GGUF | InternScience | open | reasoning | 55.74 | ● оценка |
| #68 | Agents-A1-FP8 | InternScience | open | reasoning | 55.74 | ● оценка |
| #69 | Agents-A1-Q4_K_M-GGUF | InternScience | open | reasoning | 55.74 | ● оценка |
| #70 | Agents-A1-Q8_0-GGUF | InternScience | open | reasoning | 55.74 | ● оценка |
| #71 | Qwen3.5 397B | Alibaba | open | 55.47 | ● оценка | |
| #72 | Qwen 3.6 Max (preview) | Alibaba | api | reasoning | 55.46 | ● оценка |
| #73 | Claude Sonnet 4.6 | Anthropic | api | 55.29 | ● подтверждено | |
| #74 | MiMo-V2.5-Pro | Xiaomi | api | reasoning | 55.12 | ● оценка |
| #75 | MiMo-V2-Pro | Xiaomi | api | reasoning | 54.87 | ● оценка |
| #76 | MiMo-V2.6-Flash | Xiaomi | open | reasoning | 54.78 | ● оценка |
| #77 | Apodex 1.1 | Apodex | api | reasoning | 54.57 | ● оценка |
| #78 | Gemma 4 31B | open | reasoning | 54.39 | ● оценка | |
| #79 | GLM-5 | Z.AI | open | 54.32 | ● подтверждено | |
| #80 | Qwen3.8-27B | Alibaba | open | reasoning | 54.26 | ● подтверждено |
| #81 | Gemini 3 Flash | api | 54.12 | ● подтверждено | ||
| #82 | Ternary Bonsai 2 27B | Prism ML | open | reasoning | 53.95 | ● оценка |
| #83 | Qwen3.7 Plus | Alibaba | api | reasoning | 53.79 | ● оценка |
| #84 | Kimi K2.7 Code | Moonshot AI | open | reasoning | 53.59 | ● оценка |
| #85 | Qwen3.6 Plus | Alibaba | api | reasoning | 53.54 | ● подтверждено |
| #86 | GPT-5.1-Codex | OpenAI | api | reasoning | 53.43 | ● оценка |
| #87 | Claude Opus 4.5 Thinking | Anthropic | api | reasoning | 53.4 | ● оценка |
| #88 | Claude Opus 4.5 | Anthropic | api | 53.35 | ● подтверждено | |
| #89 | MAI-Thinking-1 | Microsoft | api | reasoning | 53.02 | ● оценка |
| #90 | GPT-5.4 mini | OpenAI | api | reasoning | 52.98 | ● подтверждено |
| #91 | Hy3 Preview | Tencent | open | reasoning | 52.95 | ● оценка |
| #92 | GLM-5-Turbo | Z.AI | api | reasoning | 52.94 | ● оценка |
| #93 | Qwen3.5-122B-A10B | Alibaba | open | reasoning | 52.9 | ● оценка |
| #94 | Grok 4.3 | xAI | api | reasoning | 52.65 | ● подтверждено |
| #95 | Kimi K2.5 | Moonshot AI | open | 52.62 | ● подтверждено | |
| #96 | Mistral Large 4 | Mistral | open | 51.46 | ● подтверждено | |
| #97 | Grok 4 | xAI | api | 51.27 | ● подтверждено | |
| #98 | Muse Spark Open | Meta | open | reasoning | 50.48 | ● оценка |
| #99 | DeepSeek V3.2 | DeepSeek | open | 49.73 | ● подтверждено | |
| #100 | Inkling-Small | Thinking Machines Lab | open | 49.5 | ● подтверждено | |
| #101 | Qwen3.5 Plus | Alibaba | api | reasoning | 49.17 | ● оценка |
| #102 | MiniMax M2.5 | MiniMax | api | 49.17 | ● подтверждено | |
| #103 | Nemotron 3.5 Super VL 120B A12B | NVIDIA | open | reasoning | 49.03 | ● оценка |
| #104 | Qwen3.5-27B | Alibaba | open | reasoning | 49.01 | ● оценка |
| #105 | MiMo-V2-Omni | Xiaomi | api | reasoning | 48.99 | ● оценка |
| #106 | Gemini 2.5 Pro | api | 48.7 | ● подтверждено | ||
| #107 | GLM-5V-Turbo | Z.AI | api | 48.6 | ● оценка | |
| #108 | MiniMax M3 | MiniMax | open | 48.57 | ● подтверждено | |
| #109 | GLM-4.7 | Z.AI | open | reasoning | 48.27 | ● подтверждено |
| #110 | Qwen3.6-27B | Alibaba | open | reasoning | 48.13 | ● оценка |
| #111 | GPT-5 mini | OpenAI | api | reasoning | 47.92 | ● подтверждено |
| #112 | Inkling | Thinking Machines Lab | open | 47.91 | ● подтверждено | |
| #113 | Nemotron 3 Ultra | NVIDIA | open | reasoning | 47.37 | ● оценка |
| #114 | Qwen3.7 Flash | Alibaba | api | reasoning | 47.33 | ● оценка |
| #115 | MiniMax M2.7 | MiniMax | open | 47.29 | ● подтверждено | |
| #116 | Claude Sonnet 4.5 | Anthropic | api | 47.18 | ● оценка | |
| #117 | Gemini 3.5 Flash-Lite | api | reasoning | 46.72 | ● подтверждено | |
| #118 | Gemini 3.1 Flash-Lite | api | 46.47 | ● оценка | ||
| #119 | GPT-5.4 nano | OpenAI | api | reasoning | 46.15 | ● подтверждено |
| #120 | Ling 3.0 Flash | InclusionAI | open | reasoning | 45.33 | ● оценка |
| #121 | GPT-5 (medium) | OpenAI | api | reasoning | 45.26 | ● оценка |
| #122 | o3 | OpenAI | api | reasoning | 45.24 | ● оценка |
| #123 | Qwen3.5-35B-A3B | Alibaba | open | reasoning | 45.15 | ● подтверждено |
| #124 | Grok 4.1 Fast (Reasoning) | xAI | api | reasoning | 44.78 | ● оценка |
| #125 | Step 3.7 Flash | StepFun | open | reasoning | 44.76 | ● оценка |
| #126 | o3-pro | OpenAI | api | reasoning | 44.71 | ● подтверждено |
| #127 | Grok 3 Mini | xAI | api | reasoning | 44.09 | ● оценка |
| #128 | Trinity-Large-Thinking | Arcee AI | open | reasoning | 43.94 | ● оценка |
| #129 | Mistral Small 4 | Mistral | open | 43.16 | ● оценка | |
| #130 | Beam 1.0 501B-A23B | Reflection AI | open | reasoning | 43.08 | ● оценка |
| #131 | Qwen3.6-35B-A3B | Alibaba | open | reasoning | 43.05 | ● оценка |
| #132 | Nemotron 3.5 Bolt 74B-A9B | NVIDIA | open | reasoning | 42.8 | ● оценка |
| #133 | Grok 4 Fast (Reasoning) | xAI | api | reasoning | 42.54 | ● оценка |
| #134 | Qwen3.5 Flash | Alibaba | api | reasoning | 42.51 | ● оценка |
| #135 | GLM-4.6 | Z.AI | open | reasoning | 41.93 | ● оценка |
| #136 | Gemma 4 26B A4B | open | reasoning | 41.8 | ● оценка | |
| #137 | Step 3.5 Flash | StepFun | open | 41.69 | ● оценка | |
| #138 | Gemini 2.5 Flash | api | 41.54 | ● подтверждено | ||
| #139 | Spark-X2.6-4B | SparkLLM | open | reasoning | 40.82 | ● оценка |
| #140 | GPT-OSS 120B | OpenAI | open | 40.74 | ● оценка | |
| #141 | Grok 3 [Beta] | xAI | api | 40.68 | ● оценка | |
| #142 | DeepSeek-R1 | DeepSeek | open | reasoning | 40.4 | ● подтверждено |
| #143 | Claude Haiku 4.5 | Anthropic | api | 39.95 | ● оценка | |
| #144 | DeepSeek V3.1 | DeepSeek | open | 39.77 | ● подтверждено | |
| #145 | Claude 4.1 Opus | Anthropic | api | 39.61 | ● оценка | |
| #146 | MiMo-V2-Flash | Xiaomi | open | reasoning | 39.3 | ● оценка |
| #147 | Qwen3 Max | Alibaba | api | reasoning | 38.9 | ● оценка |
| #148 | o3-mini | OpenAI | api | reasoning | 38.8 | ● подтверждено |
| #149 | Claude 4.1 Opus Thinking | Anthropic | api | reasoning | 38.33 | ● оценка |
| #150 | DeepSeek V3.1 (Reasoning) | DeepSeek | open | reasoning | 38.16 | ● оценка |
| #151 | GLM-4.7-Flash | Z.AI | open | reasoning | 38.09 | ● оценка |
| #152 | o1 | OpenAI | api | reasoning | 38.06 | ● оценка |
| #153 | Claude 4 Sonnet | Anthropic | api | 37.65 | ● оценка | |
| #154 | GPT-OSS 20B | OpenAI | open | 37.11 | ● оценка | |
| #155 | LFM2.5-24B-A2B | LiquidAI | open | reasoning | 36.91 | ● оценка |
| #156 | GPT-4.1 | OpenAI | api | 36.8 | ● подтверждено | |
| #157 | Granite 4.2 8B | IBM | open | reasoning | 36.78 | ● оценка |
| #158 | GPT-5 nano | OpenAI | api | reasoning | 36.37 | ● подтверждено |
| #159 | o1-pro | OpenAI | api | reasoning | 36.28 | ● оценка |
| #160 | Ornith-1.5-35B-A3B | Ornith AI | open | reasoning | 36.07 | ● оценка |
| #161 | DeepSeek V3 0324 | DeepSeek | open | 35.64 | ● подтверждено | |
| #162 | Gemma 4 12B | open | reasoning | 35.48 | ● оценка | |
| #163 | Grok 4.1 Fast | xAI | api | 35.39 | ● оценка | |
| #164 | o1-preview | OpenAI | api | reasoning | 34.96 | ● подтверждено |
| #165 | ZAYA1-8B | Zyphra | open | reasoning | 34.76 | ● оценка |
| #166 | Ornith-1.5-9B | Ornith AI | open | reasoning | 34.04 | ● оценка |
| #167 | GLM-4.5-Air | Z.AI | api | 34 | ● оценка | |
| #168 | Nemotron 3 Nano Omni 30B A3B | NVIDIA | open | reasoning | 33.93 | ● оценка |
| #169 | Mercury 2.5 | Inception | api | reasoning | 33.41 | ● оценка |
| #170 | Mistral Medium 3.5 128B | Mistral | open | reasoning | 33.37 | ● оценка |
| #171 | Grok Code Fast 1 | xAI | api | 33.1 | ● оценка | |
| #172 | Mistral Large 3 | Mistral | api | 33.06 | ● оценка | |
| #173 | Mercury 2 | Inception | api | reasoning | 32.91 | ● оценка |
| #174 | Kimi K2 | Moonshot AI | api | 32.66 | ● оценка | |
| #175 | Ling 2.6 Flash | InclusionAI | open | 32.43 | ● оценка | |
| #176 | Muse Glimmer 30B | Meta | open | reasoning | 31.68 | ● оценка |
| #177 | DeepSeek V3 | DeepSeek | open | 30.85 | ● подтверждено | |
| #178 | Sarvam 105B | Sarvam | open | reasoning | 30.79 | ● оценка |
| #179 | DeepSeek R1 Distill Qwen 32B | DeepSeek | open | reasoning | 30.21 | ● оценка |
| #180 | Solar Pro 3 | Upstage | api | reasoning | 30.03 | ● оценка |
| #181 | Celeris-1 | Celeris | api | 29.46 | ● оценка | |
| #182 | GPT-4o | OpenAI | api | 29.32 | ● оценка | |
| #183 | Command A+ | Cohere | open | reasoning | 28.64 | ● оценка |
| #184 | Qwen2.5-72B | Alibaba | open | 28.5 | ● оценка | |
| #185 | GPT-4.1 mini | OpenAI | api | 28.45 | ● оценка | |
| #186 | Mistral Medium 3 | Mistral | api | 27.96 | ● оценка | |
| #187 | Llama 3.1 405B | Meta | open | 27.96 | ● оценка | |
| #188 | Mistral Large 2 | Mistral | api | 27.89 | ● оценка | |
| #189 | Llama 4 Maverick | Meta | open | 27.75 | ● оценка | |
| #190 | Gemma 3 27B | open | 27.62 | ● подтверждено | ||
| #191 | Claude 3.5 Sonnet | Anthropic | api | 27.62 | ● оценка | |
| #192 | Claude 3 Opus | Anthropic | api | 27.43 | ● подтверждено | |
| #193 | Gemini 1.5 Pro | api | 27.16 | ● оценка | ||
| #194 | Llama 4 Scout | Meta | open | 26.83 | ● оценка | |
| #195 | Gemma 4 E4B | open | reasoning | 26.58 | ● оценка | |
| #196 | MiniMax M1 80k | MiniMax | api | 26.5 | ● оценка | |
| #197 | Nemotron 3.5 Lightning 30B A3B NVFP4 | NVIDIA | open | reasoning | 26.37 | ● оценка |
| #198 | Qwen2.5 Coder 32B Instruct | Alibaba | open | 26.14 | ● оценка | |
| #199 | GPT-4o mini | OpenAI | api | 25.73 | ● подтверждено | |
| #200 | LFM2.5-2.6B | LiquidAI | open | reasoning | 25.19 | ● оценка |
| #201 | LFM2.5-8B-A1B | LiquidAI | open | reasoning | 25.05 | ● оценка |
| #202 | Sarvam 30B | Sarvam | open | reasoning | 24.21 | ● оценка |
| #203 | GPT-4.1 nano | OpenAI | api | 23.84 | ● оценка | |
| #204 | Phi-4 | Microsoft | open | 23.01 | ● оценка | |
| #205 | Nova Pro | Amazon | api | 21.19 | ● оценка | |
| #206 | LFM3-1.7B | LiquidAI | open | reasoning | 21.17 | ● оценка |
| #207 | Llama 3 70B | Meta | open | 21.09 | ● оценка | |
| #208 | GPT-4 Turbo | OpenAI | api | 20.45 | ● подтверждено | |
| #209 | Mixtral 8x22B Instruct v0.1 | Mistral | open | 18.82 | ● оценка | |
| #210 | Claude 3 Haiku | Anthropic | api | 13.2 | ● подтверждено | |
| #211 | Gemini 1.0 Pro | api | 11 | ● подтверждено | ||
| #212 | Nemotron-4 15B | NVIDIA | open | 9.89 | ● оценка | |
| #213 | Mistral 7B v0.3 | Mistral | open | 7.16 | ● оценка | |
| #214 | MiniCPM5-1B | OpenBMB | open | reasoning | 5.67 | ● оценка |
Рейтинги, цены и даты релизов на этом сайте берутся из открытых данных BenchLM.ai — сводного рейтинга языковых моделей, собирающего результаты из публикуемых бенчмарков.
Сайт агрегирует эти данные автоматически раз в сутки. Названия моделей, баллы и цены всегда сверяйте по первоисточникам провайдеров — ссылки есть на страницах моделей.
Сводный счёт модели — по шкале 0–100. Оценки опираются на публикуемые прогоны бенчмарков (SWE-bench, GPQA, MMLU-Pro, LiveCodeBench и другие), сгруппированные по задачам.