88Z.ai / RANKINGS

AI model value leaderboard

Quality versus our CNY list price when available, or Artificial Analysis’s published cost per task. Prices and benchmark scores come from different sources and should be read separately.

Last collected by this site: 2026-11-16 06:40 (UTC+8, Asia/Shanghai). Source publication dates are in the table or the citations below.

Quality versus our CNY list price when available, or Artificial Analysis’s published cost per task. Prices and benchmark scores come from different sources and should be read separately.

116 model families (best published variant per family)

Click a column header to sort. The table is fully server-rendered and remains readable without JavaScript.

#ModelValue scoreQualityOur blended CNY price (3:1)AA cost / task (USD)
1Gemini 4 CarbonGoogle96.896.0¥34.4—
2GPT 6.1 TerraOpenAI93.889.5¥17.2—
3GPT 6.1 SolOpenAI93.791.6¥34.4—
4Gemini 4 Hydrogen (High)Google93.690.4¥25.8—
5Muse Spark 1.3 (Max)Meta93.689.3¥17.2—
6Claude Sonnet 5.5 (xHigh)Anthropic93.491.1¥34.4—
7Claude Opus 5.5Anthropic93.195.0¥68.8—
8GPT 5 (High)OpenAI92.388.9¥29.56—
9Gemini 4 ArgonGoogle91.989.0¥34.4—
10Gemini 3.7 Flash (High)Google91.785.9¥12.9—
11Kimi K3 (Max)Moonshot91.488.4¥35.64—
12GLM 5.2 (Max)zai90.884.9¥14.67—
13Muse Spark 1.1Meta90.084.0¥17.2—
14Gemini 3.6 Flash (High)Google89.883.2¥12.9—
15Muse Spark 1.2 (xHigh)Meta89.883.8¥17.2—
16GLM 5Z-AI89.382.5¥13.33—
17GPT 6 Sol (Max)OpenAI89.084.7¥34.4—
18DeepSeek R1 0528DeepSeek88.680.7¥7.85—
19Grok 4.6 (High)xAI88.382.6¥25.8—
20Claude Opus 5 (Max)Anthropic88.390.2¥86—
21Hy3--87.878.8¥1.9947.8
22Mimo v2.5 Proxiaomi87.678.9¥4.68—
23Grok 4.5xAI87.581.5¥25.8—
24Claude Haiku 5.5 (High)Anthropic87.478.3¥1.72—
25Kimi k2.6Moonshot87.479.5¥11.52—
26Gemini 3.5 Flash (High)Google87.181.3¥29.03—
27Solar Pro 4$0.9687.177.9¥2.2631.39
28GLM 5.1zai87.079.8¥18.49—
29DeepSeek v4.1 Flash 1017DeepSeek86.977.6¥2.26—
30Claude Sonnet 5 (High)Anthropic86.180.5¥34.4—
31DeepSeek v3.2DeepSeek85.976.2¥2.49—
32Mimo v2.6 Flash小米集团85.976.0¥1.5139.29
33GPT 5.4OpenAI85.376.9¥14.51—
34Claude Opus 4.6 (High)Anthropic85.385.9¥86—
35Claude Opus 4.7 (High)Anthropic85.385.9¥86—
36DeepSeek V4 ProDeepSeek85.376.1¥9.03—
37GLM 4.5Z.ai85.376.0¥8.6—
38Gemini 3.1 Pro Preview--85.279.8¥38.715.56
39Claude Fable 5.1Anthropic85.196.5¥172—
40GPT 6 AstraOpenAI85.196.5¥172—
41GPT 5.2OpenAI85.079.8¥41.39—
42qwen3.6 PlusAlibaba85.075.3¥6.29—
43Kimi k2.5 (Thinking)Moonshot84.875.4¥8.6—
44Claude Sonnet 4.6Anthropic84.880.8¥51.6—
45MiniMax m2.5MiniMax84.674.6¥4.48—
46Beam 1.0 501b A23breflection84.273.9¥4.09—
47Claude Fable 5.2Anthropic84.190.6¥137.6—
48GPT 5.1OpenAI84.176.9¥29.56—
49Claude Opus 4.8 (High)Anthropic83.883.7¥86—
50GLM 4.7zai83.873.6¥6.34—
51Mimo v2.5xiaomi83.473.2¥6.88—
52GPT 5.6 TerraOpenAI83.477.1¥38.7—
53GLM 4.6Z.ai83.373.1¥7.52—
54qwen3.5 397b A17bAlibaba82.972.7¥9.35—
55DeepSeek V3 0324DeepSeek82.871.8¥4.32—
56GPT 5.4 Mini (High)OpenAI82.773.0¥14.51—
57Gemma 4 26b A4bGoogle82.671.1¥0.92—
58GLM 5.3 FlashZhipu82.170.4¥0.43102.11
59qwen3.8 Max 0902阿里巴巴82.173.5¥25.8111.02
60Gemini 2.5 ProGoogle81.372.9¥29.56—
61MiniMax m2.7MiniMax81.269.4¥3.16—
62Grok 4.3xAI81.070.4¥13.44—
63Claude Fable 5 (High)Anthropic80.189.2¥172—
64qwen3.5 122b A10bAlibaba80.168.2¥6.15—
65Mimo v2.6 Pro小米集团80.067.9¥4.6843.57
66Hy3 PreviewTencent79.967.5¥2.45—
67GLM 5.3 (Max)$5.0179.967.5¥3.4663.06
68qwen3.5 27bAlibaba79.467.3¥7.27—
69Grok 4.7xAI79.069.1¥25.8—
70Gemini 3.1 Flash Lite PreviewGoogle78.766.0¥4.84—
71qwen3.8 27bAlibaba78.665.8¥5.0—
72DeepSeek v3.2 ExpDeepSeek78.264.9¥2.62—
73Mistral Large 4 0Mistral77.765.1¥9.03—
74Gemini 2.5 FlashGoogle77.464.5¥7.31—
75Inkling 2 Small$1.1877.263.7¥4.0968.69
76Qwen3 32bAlibaba76.662.4¥1.12—
77Kimi k3.1 ProMoonshot76.564.6¥19.61178.9
78Gemini 3.8 Flash (High)$7.7775.862.9¥14.1919.58
79qwen3.5 FlashAlibaba75.761.1¥0.98—
80qwen3.5 35b A3bAlibaba75.661.2¥3.05—
81GPT 6 LunaOpenAI74.259.0¥1.72—
82GPT Oss 120bOpenAI72.857.1¥2.26—
83Muse Wonder (Max)$6.9572.364.7¥68.824.92
84Inklingthinky70.455.2¥15.16—
85Mistral Large 4Mistral70.154.0¥9.03—
86MiniMax M3稀宇科技69.752.8¥4.5279.04
87Step 5 Preview阶跃星辰69.553.4¥12.2523.76
88Gemma 3 27b ItGoogle68.150.1¥1.27—
89GPT 4o 2024 08 06OpenAI67.253.4¥37.63—
90Gemini 3.5 Flash Lite$0.6966.348.2¥7.314.91
91MiniMax m3.1 Flash稀宇科技65.446.3¥2.2650.19
92Inkling Small$0.2065.346.6¥5.4815.14
93GPT 4o Mini 2024 07 18OpenAI63.843.9¥2.26—
94Gemma 4 31bGoogle62.742.2¥2.15—
95qwen2.5 Coder 32b InstructAlibaba61.841.5¥6.41—
96Mistral Medium 3.5$1.0461.443.5¥25.823.91
97Granite 4.2 8b$0.040559.938.0¥0.7162.49
98Muse Spark OpenMeta56.635.3¥17.2—
99Agnes 2.5 Pro Theta$0.5749.523.5—5.83
100qwen3.8 Flash Next$0.5748.231.9—87.24
101DeepSeek V4 Pro 0813 (Max)$1.7745.418.1¥10.2651.48
102Apodex 1.1$0.7644.017.2—18.94
103Tencent Hy4$1.0343.114.1¥5.38115.63
104Claude Fable 5.2 Max Default Fallback$26.5441.744.5—262.07
105Nemotron 3.5 Super Vl$0.5637.86¥2.0417.14
106Nemotron 3.5 Lightning$0.3136.54¥0.8813.34
107Nemotron 3 Ultra 550b A55b Reasoning$2.2136.34.2—4.93
108Muse Ember (Max)$0.1436.03.3¥1.725.17
109Command A+$0.0035.63.2¥5.1618.3
110North Mini Code$0.0035.42.9—5.46
111K2 Horizon 375b A23b$0.0035.34.3—17.71
112Muse Glimmer (High)$0.1835.15.2—26.54
113Beam 1.0$0.2834.72.1—6.7
114Mistral Small 4 Reasoning$0.039933.91.7—12.91
115Claude Sonnet 5.5 Max Default Fallback$10.0231.033.3—296.6
116Claude Opus 5.5 Max Default Fallback$10.0416.325.4—404.53

The value score combines normalized quality (65%) and normalized cost (35%, lower is better). Site prices are CNY point list rates; AA costs are a third-party USD-per-task figure and are not the same thing.

Sources

FAQ

Where do these scores come from?
Every score comes from a public benchmark or official machine-readable file, with a source link and date. This site does not invent scores, and missing values are not treated as zero.
How is the overall score calculated?
Each source is min-max normalized within itself, then weighted: Artificial Analysis 28%, LMArena Text Arena 28%, LiveBench 22%, coding composite 15%, reasoning composite 7%. Missing sources are dropped and the rest are renormalized. A model needs at least two headline sources (or three partials) to enter the overall board. When several variants of one family exist, only the best overall variant is shown.
Are site prices part of the benchmarks?
No. Benchmark scores come from third-party public boards. CNY prices are this site’s catalogue reference rates (CNY 1 = 1,000 points) and are settled on actual usage. They are shown separately so billing is not confused with quality.
How often is this updated?
A collector runs automatically every day. The upstream boards may not publish a new snapshot daily, so each page shows both this site’s collection time and the source publication date.