模型榜
数据来源:AIHOT
专家难题、研究级数学与物理、逻辑、棋类、做实验反推规律与科学工作流,看模型能不能想明白新问题。AIHOT 评分把各家难度不同的评测折算成同一种分数:50 分是参照模型的平均水平,每 15 分约是一个标准差;两个模型相差 3 分以上,先后基本不会变。评分不是正确率。分类评分只看这个领域,只列在这个领域测得够多的模型。
15 项评测 · 7 家机构 · 更新 2026-11-18 20:05 北京时间
国产由厂商推断,开源依据模型页的权重链接;暂不排名的模型缺少权重标记,开源筛选时不展示。价格单位:¥ / 百万 Token。
| 名次 | 模型 | 上线日期 | 评测 | 缓存价 | 输入价 | 输出价 | 评分 |
|---|---|---|---|---|---|---|---|
| 1 | Claude Fable 5.2Anthropic | 2026-11-09 | 11 项评测 | ¥1.33 | ¥53.33 | ¥266.64 | 82.1 |
| 2 | gpt-6.2-astraOpenAI | 2026-11-17 | 9 项评测 | ¥2.67 | ¥53.33 | ¥266.64 | 81.1 |
| 3 | Gemini 4 CarbonGoogle | 2026-10-28 | 6 项评测 | ¥1.33 | ¥26.66 | ¥133.32 | 78.9 |
| 4 | GPT-6 AstraOpenAI | 2026-09-03 | 15 项评测 | ¥6.67 | ¥66.66 | ¥333.31 | 76.8 |
| 5 | Claude Opus 5.5Anthropic | 2026-09-22 | 15 项评测 | ¥1.33 | ¥26.66 | ¥133.32 | 75.4 |
| 6 | GPT-6.1 SolOpenAI | 2026-09-29 | 15 项评测 | ¥0.67 | ¥13.33 | ¥66.66 | 75.3 |
| 7 | gpt-6.2-solOpenAI | 2026-11-17 | 7 项评测 | ¥0.67 | ¥13.33 | ¥66.66 | 74.8 |
| 8 | Gemini 4 HydrogenGoogle | 2026-11-12 | 3 项评测 | ¥0.5 | ¥10 | ¥50 | 74.6 |
| 9 | Claude Fable 5.1Anthropic | 2026-09-01 | 12 项评测 | ¥1.67 | ¥66.66 | ¥333.31 | 73.3 |
| 10 | GPT-6.1 TerraOpenAI | 2026-10-26 | 6 项评测 | ¥0.33 | ¥6.67 | ¥33.33 | 73.3 |
| 11 | Claude Sonnet 5.5Anthropic | 2026-09-28 | 15 项评测 | ¥0.67 | ¥13.33 | ¥66.66 | 73.1 |
| 12 | Gemini 4 ArgonGoogle | 2026-09-30 | 14 项评测 | ¥1.33 | ¥26.66 | ¥133.32 | 72.5 |
| 13 | Claude Opus 5Anthropic | 2026-07-24 | 12 项评测 | ¥3.33 | ¥33.33 | ¥166.65 | 70.7 |
| 14 | GPT-5.6 SolOpenAI | 2026-07-09 | 12 项评测 | ¥3.33 | ¥33.33 | ¥199.98 | 70.2 |
| 15 | Claude Fable 5Anthropic | 2026-06-09 | 8 项评测 | ¥6.67 | ¥66.66 | ¥333.31 | 69.5 |
| 16 | GPT-6 SolOpenAI | 2026-09-22 | 12 项评测 | ¥1.33 | ¥13.33 | ¥66.66 | 69.5 |
| 17 | Kimi K3.1 ProMoonshot AI | 2026-10-23 | 10 项评测 | ¥1.3 | ¥13 | ¥68 | 69.4 |
| 18 | Muse WonderMeta | 2026-10-12 | 11 项评测 | ¥1.33 | ¥26.66 | ¥133.32 | 67.7 |
| 19 | GPT-5.5OpenAI | 2026-04-23 | 7 项评测 | ¥3.33 | ¥33.33 | ¥199.98 | 67.7 |
| 20 | Qwen4 MaxAlibaba | 2026-11-16 | 5 项评测 | ¥0.5 | ¥10 | ¥30 | 67.0 |
| 21 | GPT-5.6 TerraOpenAI | 2026-07-09 | 10 项评测 | ¥1.33 | ¥13.33 | ¥79.99 | 65.7 |
| 22 | Muse EmberMeta | 2026-11-09 | 3 项评测 | ¥0.0267 | ¥0.67 | ¥3.33 | 65.5 |
| 23 | Gemini 3.8 FlashGoogle | 2026-09-02 | 15 项评测 | ¥1 | ¥10 | ¥50 | 65.3 |
| 24 | Hy4Tencent | 2026-11-04 | 4 项评测 | ¥0.3 | ¥6 | ¥18 | 64.9 |
| 25 | Muse Spark 1.3Meta | 2026-09-02 | 10 项评测 | ¥1 | ¥8.33 | ¥28.33 | 64.6 |
| 26 | DeepSeek V4.1 Flash 1017DeepSeek | 2026-10-17 | 11 项评测 | ¥0.02 | ¥1 | ¥4 | 64.6 |
| 27 | Hy 4 PreviewTencent | 2026-08-28 | 3 项评测 | ¥0.3 | ¥6 | ¥18 | 64.3 |
| 28 | Gemini 3.7 FlashGoogle | 2026-08-13 | 8 项评测 | ¥1 | ¥10 | ¥50 | 63.8 |
| 29 | Qwen3.8 MaxAlibaba | 2026-07-19 | 11 项评测 | ¥1.5 | ¥12 | ¥36 | 63.6 |
| 30 | Claude Opus 4.8Anthropic | 2026-05-28 | 8 项评测 | ¥3.33 | ¥33.33 | ¥166.65 | 63.4 |
| 31 | Claude Haiku 5.5Anthropic | 2026-10-07 | 13 项评测 | ¥0.0667 | ¥0.67 | ¥3.33 | 63.3 |
| 32 | Grok 4.6xAI | 2026-08-12 | 12 项评测 | ¥3.33 | ¥13.33 | ¥40 | 62.7 |
| 33 | Inkling 2 SmallThinking Machines Lab | 2026-11-05 | 7 项评测 | ¥0.2 | ¥2 | ¥6.67 | 62.7 |
| 34 | Kimi K3Moonshot AI | 2026-07-16 | 13 项评测 | ¥2 | ¥20 | ¥100 | 62.6 |
| 35 | GPT-5.4OpenAI | 2026-03-05 | 4 项评测 | ¥1.67 | ¥16.67 | ¥99.99 | 62.3 |
| 36 | Grok 4.7xAI | 2026-09-21 | 15 项评测 | ¥3.33 | ¥13.33 | ¥40 | 61.6 |
| 37 | GLM-5.3Z.ai | 2026-08-18 | 13 项评测 | ¥2 | ¥8 | ¥28 | 61.5 |
| 38 | Qwen3.8 Max 0902Alibaba | 2026-09-01 | 6 项评测 | ¥1.5 | ¥12 | ¥36 | 61.1 |
| 39 | Claude Opus 4.7Anthropic | 2026-04-16 | 7 项评测 | ¥3.33 | ¥33.33 | ¥166.65 | 60.9 |
| 40 | DeepSeek V4.1 FlashDeepSeek | 2026-09-10 | 12 项评测 | ¥0.02 | ¥1 | ¥4 | 60.8 |
| 41 | Muse Spark 1.2Meta | 2026-08-05 | 5 项评测 | ¥1 | ¥8.33 | ¥28.33 | 60.7 |
| 42 | MiniMax-M3.1-FlashMiniMax | 2026-10-19 | 7 项评测 | ¥0.1 | ¥1 | ¥4 | 60.5 |
| 43 | Step 5 PreviewStepFun | 2026-09-18 | 10 项评测 | ¥0.7 | ¥7 | ¥20 | 60.3 |
| 44 | Claude Sonnet 5Anthropic | 2026-06-30 | 7 项评测 | ¥1.33 | ¥13.33 | ¥66.66 | 60.2 |
| 45 | DeepSeek V4 Pro 0813DeepSeek | 2026-08-13 | 13 项评测 | ¥0.15 | ¥4.5 | ¥13.5 | 60.0 |
| 46 | MiMo-V2.6-ProXiaomi | 2026-09-21 | 8 项评测 | ¥0.025 | ¥3 | ¥6 | 60.0 |
| 47 | Muse Spark 1.1Meta | 2026-07-09 | 4 项评测 | ¥1 | ¥8.33 | ¥28.33 | 59.9 |
| 48 | Gemini 3.1 Pro PreviewGoogle | 2026-02-19 | 10 项评测 | ¥1.33 | ¥13.33 | ¥79.99 | 59.7 |
| 49 | Gemini 3.5 FlashGoogle | 2026-05-19 | 8 项评测 | ¥1 | ¥10 | ¥59.99 | 58.9 |
| 50 | Claude Opus 4.6Anthropic | 2026-02-05 | 7 项评测 | ¥3.33 | ¥33.33 | ¥166.65 | 58.9 |
一般用量:缓存命中:输入:输出 = 7:2:1。 2026-11-17 · $1 = ¥6.6661
暂不排名 · 5
Muse Wonder FullMeta
70.6本轮未收录官方 API 价格或公开权重
ember-1Fireworks AI
61.8评测数量还不够
Qwen3.8 2.4 T A 95 BAlibaba
61.4评测数量还不够
60.4
Ling 3.1 FlashInclusionAI
59评测数量还不够
全部模型 · 45
| 名次 | 模型 | 上线日期 | 评测 | 缓存价 | 输入价 | 输出价 | 评分 |
|---|---|---|---|---|---|---|---|
| 51 | GPT-6 LunaOpenAI | 2026-09-22 | 13 项评测 | ¥0.0667 | ¥0.67 | ¥3.33 | 58.9 |
| 52 | Qwen3.8 Flash NextAlibaba | 2026-08-26 | 5 项评测 | ¥0.1 | ¥0.8 | ¥2.7 | 58.8 |
| 53 | GPT-5.6 LunaOpenAI | 2026-07-09 | 12 项评测 | ¥0.13 | ¥1.33 | ¥8 | 58.7 |
| 54 | Grok 4.5xAI | 2026-07-08 | 6 项评测 | ¥2 | ¥13.33 | ¥40 | 58.7 |
| 55 | DeepSeek V4 Flash 0731DeepSeek | 2026-07-31 | 9 项评测 | ¥0.05 | ¥1.5 | ¥4.5 | 58.3 |
| 56 | MiMo V 2.6 FlashXiaomi | 2026-09-21 | 7 项评测 | ¥0.02 | ¥1 | ¥2 | 58.2 |
| 57 | Muse Spark OpenMeta | 2026-10-15 | 5 项评测 | — | — | — | 57.8 |
| 58 | Gemini 3.6 FlashGoogle | 2026-07-21 | 9 项评测 | ¥1 | ¥10 | ¥50 | 57.4 |
| 59 | Qwen3.7 MaxAlibaba | 2026-05-19 | 6 项评测 | ¥2.4 | ¥12 | ¥36 | 57.0 |
| 60 | GLM-5.3 FlashZ.ai | 2026-08-26 | 13 项评测 | ¥0.23 | ¥0.8 | ¥2.8 | 56.5 |
| 61 | GLM-5.2Z.ai | 2026-06-16 | 7 项评测 | ¥2 | ¥8 | ¥28 | 54.0 |
| 62 | Mistral Large 4Mistral AI | 2026-10-06 | 6 项评测 | ¥0.47 | ¥4.67 | ¥14 | 53.6 |
| 63 | Beam 1.0Reflection | 2026-10-16 | 8 项评测 | — | — | — | 53.0 |
| 64 | DeepSeek V4 Pro PreviewDeepSeek | 2026-04-24 | 8 项评测 | ¥0.025 | ¥3 | ¥6 | 52.5 |
| 65 | Qwen3.8-27BAlibaba | 2026-08-14 | 8 项评测 | ¥0.6 | ¥3 | ¥12 | 52.2 |
| 66 | Claude Sonnet 4.6Anthropic | 2026-02-17 | 6 项评测 | ¥2 | ¥20 | ¥99.99 | 52.2 |
| 67 | Kimi K2.6Moonshot AI | 2026-04-20 | 6 项评测 | ¥1.1 | ¥6.5 | ¥27 | 52.2 |
| 68 | Kimi K2.7 CodeMoonshot AI | 2026-06-12 | 5 项评测 | ¥1.3 | ¥6.5 | ¥27 | 51.6 |
| 69 | MiniMax-M3MiniMax | 2026-06-01 | 9 项评测 | ¥0.42 | ¥2.1 | ¥8.4 | 50.8 |
| 70 | Qwen3.7 PlusAlibaba | 2026-06-01 | 4 项评测 | ¥0.4 | ¥2 | ¥8 | 49.4 |
| 71 | DeepSeek V4 Flash PreviewDeepSeek | 2026-04-24 | 3 项评测 | ¥0.02 | ¥1 | ¥2 | 49.1 |
| 72 | InklingThinking Machines Lab | 2026-07-15 | 8 项评测 | ¥1.13 | ¥6.67 | ¥27 | 48.3 |
| 73 | MiMo-V2.5-ProXiaomi | 2026-04-22 | 2 项评测 | ¥0.025 | ¥3 | ¥6 | 48.1 |
| 74 | Grok 4.3xAI | 2026-04-17 | 5 项评测 | ¥1.33 | ¥8.33 | ¥16.67 | 47.5 |
| 75 | GPT-5.4 Mini 2026-03-17OpenAI | 2026-03-17 | 7 项评测 | ¥0.5 | ¥5 | ¥30 | 47.2 |
| 76 | Inkling-SmallThinking Machines Lab | 2026-07-30 | 7 项评测 | ¥0.4 | ¥2 | ¥8 | 47.1 |
| 77 | GPT-5.4 Nano 2026-03-17OpenAI | 2026-03-17 | 7 项评测 | ¥0.13 | ¥1.33 | ¥8.33 | 46.8 |
| 78 | GLM-5.1Z.ai | 2026-04-07 | 4 项评测 | ¥1.3 | ¥6 | ¥24 | 46.1 |
| 79 | Nemotron 3 UltraNVIDIA | 2026-06-04 | 5 项评测 | — | — | — | 46.0 |
| 80 | Qwen3.6 PlusAlibaba | 2026-04-01 | 6 项评测 | ¥0.2 | ¥2 | ¥12 | 45.6 |
| 81 | Nemotron 3.5 Super VLNVIDIA | 2026-10-15 | 5 项评测 | — | — | — | 45.5 |
| 82 | MiMo-V2.5Xiaomi | 2026-04-22 | 2 项评测 | ¥0.02 | ¥1 | ¥2 | 43.5 |
| 83 | Qwen3.6-27BAlibaba | 2026-04-22 | 5 项评测 | — | ¥3 | ¥18 | 42.5 |
| 84 | MiniMax M 2.7MiniMax | 2026-03-18 | 2 项评测 | ¥0.42 | ¥2.1 | ¥8.4 | 41.2 |
| 85 | Gemini 3.5 Flash LiteGoogle | 2026-07-21 | 10 项评测 | ¥0.2 | ¥2 | ¥16.67 | 40.6 |
| 86 | Muse GlimmerMeta | 2026-08-10 | 4 项评测 | — | — | — | 39.3 |
| 87 | GPT-5 MiniOpenAI | 2025-08-07 | 6 项评测 | ¥0.17 | ¥1.67 | ¥13.33 | 38.3 |
| 88 | Gemma 4 31 BGoogle | 2026-04-02 | 4 项评测 | 免费 | 免费 | 免费 | 37.6 |
| 89 | Gemini 2.5 ProGoogle | 2025-06-17 | 4 项评测 | ¥0.83 | ¥8.33 | ¥66.66 | 34.5 |
| 90 | Gemini 3.1 Flash Lite PreviewGoogle | 2026-03-03 | 4 项评测 | ¥0.17 | ¥1.67 | ¥10 | 30.9 |
| 91 | Mistral Medium 3.5Mistral AI | 2026-04-29 | 4 项评测 | ¥1 | ¥10 | ¥50 | 29.9 |
| 92 | Claude Haiku 4.5 20251001Anthropic | 2025-10-15 | 2 项评测 | ¥0.67 | ¥6.67 | ¥33.33 | 28.5 |
| 93 | gpt-oss-120bOpenAI | 2025-08-05 | 5 项评测 | — | — | — | 25.9 |
| 94 | Granite 4.2 8 BIBM | 2026-08-25 | 4 项评测 | — | — | — | 20.8 |
| 95 | Granite 4.2 3 BIBM | 2026-08-25 | 4 项评测 | — | — | — | 10.5 |