AI中转站/模型比价

模型能力排名

基于 OpenRouter Benchmarks数据,按各能力分数从高到低展示,数值越大代表能力越优秀。

综合智能指数

指数 从高到低
排名模型供应商上下文指数
1 Claude Opus 5anthropic/claude-opus-5 anthropic 1M 63.1
2 Claude Opus 5 (batch)anthropic/claude-opus-5:batch anthropic 1M 63.1
3 Anthropic: Claude Fable 5anthropic/claude-fable-5 anthropic 1M 62.1
4 Anthropic: Claude Fable 5 (batch)anthropic/claude-fable-5:batch anthropic 1M 62.1
5 OpenAI: GPT-5.6 Solopenai/gpt-5.6-sol openai 1.1M 60.9
6 OpenAI: GPT-5.6 Sol (batch)openai/gpt-5.6-sol:batch openai 1.1M 60.9
7 SpaceXAI: Grok 4.6x-ai/grok-4.6 x-ai 500K 60.9
8 MoonshotAI: Kimi K3moonshotai/kimi-k3 moonshotai 1M 59.7
9 Qwen: Qwen3.8 Maxqwen/qwen3.8-max qwen 1M 58.1
10 Anthropic: Claude Opus 4.8anthropic/claude-opus-4.8 anthropic 1M 57.3
11 Anthropic: Claude Opus 4.8 (batch)anthropic/claude-opus-4.8:batch anthropic 1M 57.3
12 OpenAI: GPT-5.6 Terraopenai/gpt-5.6-terra openai 1.1M 56.6
13 OpenAI: GPT-5.6 Terra (batch)openai/gpt-5.6-terra:batch openai 1.1M 56.6
14 OpenAI: GPT-5.5openai/gpt-5.5 openai 1.1M 56.3
15 OpenAI: GPT-5.5 (batch)openai/gpt-5.5:batch openai 1.1M 56.3
16 Google: Gemini 3.7 Flashgoogle/gemini-3.7-flash google 1M 56
17 Google: Gemini 3.7 Flash (batch)google/gemini-3.7-flash:batch google 1M 56
18 SpaceXAI: Grok 4.5x-ai/grok-4.5 x-ai 500K 55.8
19 Anthropic: Claude Sonnet 5anthropic/claude-sonnet-5 anthropic 1M 55.3
20 Anthropic: Claude Sonnet 5 (batch)anthropic/claude-sonnet-5:batch anthropic 1M 55.3
21 Anthropic: Claude Opus 4.7anthropic/claude-opus-4.7 anthropic 1M 55
22 Anthropic: Claude Opus 4.7 (batch)anthropic/claude-opus-4.7:batch anthropic 1M 55
23 DeepSeek: DeepSeek V4 Pro 0813deepseek/deepseek-v4-pro-0813 deepseek 1M 53.2
24 Meta: Muse Spark 1.1meta/muse-spark-1.1 meta 1M 53.2
25 OpenAI: GPT-5.4openai/gpt-5.4 openai 1.1M 53.1
26 OpenAI: GPT-5.4 (batch)openai/gpt-5.4:batch openai 1.1M 53.1
27 Z.ai: GLM 5.2z-ai/glm-5.2 z-ai 1M 52.6
28 Z.ai: GLM 5.2 (batch)z-ai/glm-5.2:batch z-ai 512K 52.6
29 OpenAI: GPT-5.6 Lunaopenai/gpt-5.6-luna openai 1.1M 52.3
30 OpenAI: GPT-5.6 Luna (batch)openai/gpt-5.6-luna:batch openai 1.1M 52.3

编码能力指数

指数 从高到低
排名模型供应商上下文指数
1 Claude Opus 5anthropic/claude-opus-5 anthropic 1M 78
2 Claude Opus 5 (batch)anthropic/claude-opus-5:batch anthropic 1M 78
3 OpenAI: GPT-5.6 Solopenai/gpt-5.6-sol openai 1.1M 77.4
4 OpenAI: GPT-5.6 Sol (batch)openai/gpt-5.6-sol:batch openai 1.1M 77.4
5 SpaceXAI: Grok 4.6x-ai/grok-4.6 x-ai 500K 76.8
6 OpenAI: GPT-5.6 Terraopenai/gpt-5.6-terra openai 1.1M 76.7
7 OpenAI: GPT-5.6 Terra (batch)openai/gpt-5.6-terra:batch openai 1.1M 76.7
8 Anthropic: Claude Fable 5anthropic/claude-fable-5 anthropic 1M 76.5
9 Anthropic: Claude Fable 5 (batch)anthropic/claude-fable-5:batch anthropic 1M 76.5
10 MoonshotAI: Kimi K3moonshotai/kimi-k3 moonshotai 1M 76.2
11 Google: Gemini 3.7 Flashgoogle/gemini-3.7-flash google 1M 76.1
12 Google: Gemini 3.7 Flash (batch)google/gemini-3.7-flash:batch google 1M 76.1
13 OpenAI: GPT-5.5openai/gpt-5.5 openai 1.1M 74.9
14 OpenAI: GPT-5.5 (batch)openai/gpt-5.5:batch openai 1.1M 74.9
15 Anthropic: Claude Opus 4.8anthropic/claude-opus-4.8 anthropic 1M 74.3
16 Anthropic: Claude Opus 4.8 (batch)anthropic/claude-opus-4.8:batch anthropic 1M 74.3
17 Anthropic: Claude Opus 4.7anthropic/claude-opus-4.7 anthropic 1M 73.6
18 Anthropic: Claude Opus 4.7 (batch)anthropic/claude-opus-4.7:batch anthropic 1M 73.6
19 SpaceXAI: Grok 4.5x-ai/grok-4.5 x-ai 500K 72.4
20 Qwen: Qwen3.8 Maxqwen/qwen3.8-max qwen 1M 71.8
21 Anthropic: Claude Sonnet 5anthropic/claude-sonnet-5 anthropic 1M 71.5
22 Anthropic: Claude Sonnet 5 (batch)anthropic/claude-sonnet-5:batch anthropic 1M 71.5
23 OpenAI: GPT-5.6 Lunaopenai/gpt-5.6-luna openai 1.1M 71.4
24 OpenAI: GPT-5.6 Luna (batch)openai/gpt-5.6-luna:batch openai 1.1M 71.4
25 Meta: Muse Spark 1.1meta/muse-spark-1.1 meta 1M 71.3
26 OpenAI: GPT-5.4openai/gpt-5.4 openai 1.1M 71.1
27 OpenAI: GPT-5.4 (batch)openai/gpt-5.4:batch openai 1.1M 71.1
28 Google: Gemini 3.5 Flashgoogle/gemini-3.5-flash google 1M 70.1
29 Google: Gemini 3.5 Flash (batch)google/gemini-3.5-flash:batch google 1M 70.1
30 Google: Gemini 3.6 Flashgoogle/gemini-3.6-flash google 1M 69.2

智能体能力指数

指数 从高到低
排名模型供应商上下文指数
1 Claude Opus 5anthropic/claude-opus-5 anthropic 1M 59.2
2 Claude Opus 5 (batch)anthropic/claude-opus-5:batch anthropic 1M 59.2
3 SpaceXAI: Grok 4.6x-ai/grok-4.6 x-ai 500K 58.7
4 Qwen: Qwen3.8 Maxqwen/qwen3.8-max qwen 1M 58.4
5 OpenAI: GPT-5.6 Solopenai/gpt-5.6-sol openai 1.1M 57.8
6 OpenAI: GPT-5.6 Sol (batch)openai/gpt-5.6-sol:batch openai 1.1M 57.8
7 Anthropic: Claude Fable 5anthropic/claude-fable-5 anthropic 1M 56.6
8 Anthropic: Claude Fable 5 (batch)anthropic/claude-fable-5:batch anthropic 1M 56.6
9 MoonshotAI: Kimi K3moonshotai/kimi-k3 moonshotai 1M 54.3
10 OpenAI: GPT-5.6 Terraopenai/gpt-5.6-terra openai 1.1M 50.2
11 OpenAI: GPT-5.6 Terra (batch)openai/gpt-5.6-terra:batch openai 1.1M 50.2
12 Anthropic: Claude Sonnet 5anthropic/claude-sonnet-5 anthropic 1M 49.7
13 Anthropic: Claude Sonnet 5 (batch)anthropic/claude-sonnet-5:batch anthropic 1M 49.7
14 DeepSeek: DeepSeek V4 Pro 0813deepseek/deepseek-v4-pro-0813 deepseek 1M 49.6
15 Anthropic: Claude Opus 4.8anthropic/claude-opus-4.8 anthropic 1M 49.4
16 Anthropic: Claude Opus 4.8 (batch)anthropic/claude-opus-4.8:batch anthropic 1M 49.4
17 SpaceXAI: Grok 4.5x-ai/grok-4.5 x-ai 500K 48.9
18 DeepSeek: DeepSeek V4 Flash 0731deepseek/deepseek-v4-flash-0731 deepseek 1M 48.4
19 OpenAI: GPT-5.5openai/gpt-5.5 openai 1.1M 47.4
20 OpenAI: GPT-5.5 (batch)openai/gpt-5.5:batch openai 1.1M 47.4
21 OpenAI: GPT-5.6 Lunaopenai/gpt-5.6-luna openai 1.1M 46.9
22 OpenAI: GPT-5.6 Luna (batch)openai/gpt-5.6-luna:batch openai 1.1M 46.9
23 Anthropic: Claude Opus 4.7anthropic/claude-opus-4.7 anthropic 1M 46.3
24 Anthropic: Claude Opus 4.7 (batch)anthropic/claude-opus-4.7:batch anthropic 1M 46.3
25 Z.ai: GLM 5.2z-ai/glm-5.2 z-ai 1M 45.7
26 Z.ai: GLM 5.2 (batch)z-ai/glm-5.2:batch z-ai 512K 45.7
27 Google: Gemini 3.7 Flashgoogle/gemini-3.7-flash google 1M 45.1
28 Google: Gemini 3.7 Flash (batch)google/gemini-3.7-flash:batch google 1M 45.1
29 OpenAI: GPT-5.4openai/gpt-5.4 openai 1.1M 44.2
30 OpenAI: GPT-5.4 (batch)openai/gpt-5.4:batch openai 1.1M 44.2

Agent 场景 / agenticgamedev

Elo 从高到低
排名模型供应商上下文Elo胜率源排名
1 Anthropic: Claude Fable 5anthropic/claude-fable-5 anthropic 1M 1298 65.1% 1
2 Anthropic: Claude Fable 5 (batch)anthropic/claude-fable-5:batch anthropic 1M 1298 65.1% 1
3 Anthropic: Claude Opus 4.8anthropic/claude-opus-4.8 anthropic 1M 1258 61.5% 2
4 Anthropic: Claude Opus 4.8 (batch)anthropic/claude-opus-4.8:batch anthropic 1M 1258 61.5% 2
5 Anthropic: Claude Opus 4.7anthropic/claude-opus-4.7 anthropic 1M 1251 61.4% 3
6 Anthropic: Claude Opus 4.7 (batch)anthropic/claude-opus-4.7:batch anthropic 1M 1251 61.4% 3
7 Anthropic: Claude Sonnet 5anthropic/claude-sonnet-5 anthropic 1M 1236 55.7% 4
8 Anthropic: Claude Sonnet 5 (batch)anthropic/claude-sonnet-5:batch anthropic 1M 1236 55.7% 4
9 SpaceXAI: Grok 4.5x-ai/grok-4.5 x-ai 500K 1221 56.6% 5
10 Google: Gemini 3.6 Flashgoogle/gemini-3.6-flash google 1M 1204 53.7% 6
11 Google: Gemini 3.6 Flash (batch)google/gemini-3.6-flash:batch google 1M 1204 53.7% 6
12 MiniMax: MiniMax M3minimax/minimax-m3 minimax 1M 1195 52% 7
13 MiniMax: MiniMax M3 (batch)minimax/minimax-m3:batch minimax 524.3K 1195 52% 7
14 Meta: Muse Spark 1.1meta/muse-spark-1.1 meta 1M 1190 48.2% 8
15 Anthropic: Claude Sonnet 4.6anthropic/claude-sonnet-4.6 anthropic 1M 1187 52.2% 9
16 Anthropic: Claude Sonnet 4.6 (batch)anthropic/claude-sonnet-4.6:batch anthropic 1M 1187 52.2% 9
17 Z.ai: GLM 5.2z-ai/glm-5.2 z-ai 1M 1185 48.8% 10
18 Z.ai: GLM 5.2 (batch)z-ai/glm-5.2:batch z-ai 512K 1185 48.8% 10
19 Google: Gemini 3.5 Flashgoogle/gemini-3.5-flash google 1M 1184 54% 11
20 Google: Gemini 3.5 Flash (batch)google/gemini-3.5-flash:batch google 1M 1184 54% 11
21 OpenAI: GPT-5.5openai/gpt-5.5 openai 1.1M 1182 51.4% 12
22 OpenAI: GPT-5.5 (batch)openai/gpt-5.5:batch openai 1.1M 1182 51.4% 12
23 Qwen: Qwen3.7 Maxqwen/qwen3.7-max qwen 1M 1176 48.3% 13
24 Z.ai: GLM 5.1z-ai/glm-5.1 z-ai 204.8K 1174 50.7% 14
25 MoonshotAI: Kimi K2.6moonshotai/kimi-k2.6 moonshotai 262.1K 1148 47.9% 15
26 MoonshotAI: Kimi K2.7 Codemoonshotai/kimi-k2.7-code moonshotai 262.1K 1137 43% 16
27 MoonshotAI: Kimi K2.7 Code (batch)moonshotai/kimi-k2.7-code:batch moonshotai 262.1K 1137 43% 16
28 Google: Gemini 3.1 Pro Previewgoogle/gemini-3.1-pro-preview google 1M 1121 44% 17
29 Google: Gemini 3.1 Pro Preview (batch)google/gemini-3.1-pro-preview:batch google 1M 1121 44% 17
30 Z.ai: GLM 5V Turboz-ai/glm-5v-turbo z-ai 202.8K 1104 41.4% 18

Agent 场景 / agentichtmlslides

Elo 从高到低
排名模型供应商上下文Elo胜率源排名
1 Anthropic: Claude Fable 5anthropic/claude-fable-5 anthropic 1M 1254 59.4% 1
2 Anthropic: Claude Fable 5 (batch)anthropic/claude-fable-5:batch anthropic 1M 1254 59.4% 1
3 MoonshotAI: Kimi K2.6moonshotai/kimi-k2.6 moonshotai 262.1K 1248 59% 2
4 Anthropic: Claude Opus 4.7anthropic/claude-opus-4.7 anthropic 1M 1243 58% 3
5 Anthropic: Claude Opus 4.7 (batch)anthropic/claude-opus-4.7:batch anthropic 1M 1243 58% 3
6 Anthropic: Claude Opus 4.8anthropic/claude-opus-4.8 anthropic 1M 1227 55.6% 4
7 Anthropic: Claude Opus 4.8 (batch)anthropic/claude-opus-4.8:batch anthropic 1M 1227 55.6% 4
8 Google: Gemini 3.1 Pro Previewgoogle/gemini-3.1-pro-preview google 1M 1226 55.8% 5
9 Google: Gemini 3.1 Pro Preview (batch)google/gemini-3.1-pro-preview:batch google 1M 1226 55.8% 5
10 Z.ai: GLM 5.1z-ai/glm-5.1 z-ai 204.8K 1205 52% 6
11 Google: Gemini 3.5 Flashgoogle/gemini-3.5-flash google 1M 1162 45.8% 7
12 Google: Gemini 3.5 Flash (batch)google/gemini-3.5-flash:batch google 1M 1162 45.8% 7
13 Z.ai: GLM 5V Turboz-ai/glm-5v-turbo z-ai 202.8K 1134 41.7% 8
14 OpenAI: GPT-5.5openai/gpt-5.5 openai 1.1M 1084 34.2% 9
15 OpenAI: GPT-5.5 (batch)openai/gpt-5.5:batch openai 1.1M 1084 34.2% 9
16 SpaceXAI: Grok 4.3x-ai/grok-4.3 x-ai 1M 1066 31.8% 10

Agent 场景 / agenticslides

Elo 从高到低
排名模型供应商上下文Elo胜率源排名
1 Anthropic: Claude Opus 4.7anthropic/claude-opus-4.7 anthropic 1M 1334 64.7% 1
2 Anthropic: Claude Opus 4.7 (batch)anthropic/claude-opus-4.7:batch anthropic 1M 1334 64.7% 1
3 Anthropic: Claude Opus 4.8anthropic/claude-opus-4.8 anthropic 1M 1294 64.8% 2
4 Anthropic: Claude Opus 4.8 (batch)anthropic/claude-opus-4.8:batch anthropic 1M 1294 64.8% 2
5 Z.ai: GLM 5.1z-ai/glm-5.1 z-ai 204.8K 1245 54.4% 3
6 Google: Gemini 3.5 Flashgoogle/gemini-3.5-flash google 1M 1244 57.5% 4
7 Google: Gemini 3.5 Flash (batch)google/gemini-3.5-flash:batch google 1M 1244 57.5% 4
8 MoonshotAI: Kimi K2.6moonshotai/kimi-k2.6 moonshotai 262.1K 1187 45.8% 5
9 Z.ai: GLM 5V Turboz-ai/glm-5v-turbo z-ai 202.8K 1171 51.6% 6
10 OpenAI: GPT-5.5openai/gpt-5.5 openai 1.1M 1150 43.5% 7
11 OpenAI: GPT-5.5 (batch)openai/gpt-5.5:batch openai 1.1M 1150 43.5% 7
12 Google: Gemini 3.1 Pro Previewgoogle/gemini-3.1-pro-preview google 1M 1112 33.8% 8
13 Google: Gemini 3.1 Pro Preview (batch)google/gemini-3.1-pro-preview:batch google 1M 1112 33.8% 8
14 Google: Gemini 3 Flash Previewgoogle/gemini-3-flash-preview google 1M 1073 39.3% 9
15 Google: Gemini 3 Flash Preview (batch)google/gemini-3-flash-preview:batch google 1M 1073 39.3% 9
16 SpaceXAI: Grok 4.3x-ai/grok-4.3 x-ai 1M 1072 31.9% 10

Agent 场景 / agenticslides(html)

Elo 从高到低
排名模型供应商上下文Elo胜率源排名
1 Anthropic: Claude Fable 5anthropic/claude-fable-5 anthropic 1M 1252 59.5% 1
2 Anthropic: Claude Fable 5 (batch)anthropic/claude-fable-5:batch anthropic 1M 1252 59.5% 1
3 MoonshotAI: Kimi K2.6moonshotai/kimi-k2.6 moonshotai 262.1K 1252 59.2% 2
4 Anthropic: Claude Opus 4.7anthropic/claude-opus-4.7 anthropic 1M 1242 57.8% 3
5 Anthropic: Claude Opus 4.7 (batch)anthropic/claude-opus-4.7:batch anthropic 1M 1242 57.8% 3
6 Anthropic: Claude Opus 4.8anthropic/claude-opus-4.8 anthropic 1M 1230 56% 4
7 Anthropic: Claude Opus 4.8 (batch)anthropic/claude-opus-4.8:batch anthropic 1M 1230 56% 4
8 Google: Gemini 3.1 Pro Previewgoogle/gemini-3.1-pro-preview google 1M 1219 54.4% 5
9 Google: Gemini 3.1 Pro Preview (batch)google/gemini-3.1-pro-preview:batch google 1M 1219 54.4% 5
10 Z.ai: GLM 5.1z-ai/glm-5.1 z-ai 204.8K 1204 51.8% 6
11 Google: Gemini 3.5 Flashgoogle/gemini-3.5-flash google 1M 1162 45.7% 7
12 Google: Gemini 3.5 Flash (batch)google/gemini-3.5-flash:batch google 1M 1162 45.7% 7
13 Z.ai: GLM 5V Turboz-ai/glm-5v-turbo z-ai 202.8K 1137 41.8% 8
14 OpenAI: GPT-5.5openai/gpt-5.5 openai 1.1M 1077 33.2% 9
15 OpenAI: GPT-5.5 (batch)openai/gpt-5.5:batch openai 1.1M 1077 33.2% 9
16 SpaceXAI: Grok 4.3x-ai/grok-4.3 x-ai 1M 1066 31.7% 10

Agent 场景 / agenticslides(python-pptx)

Elo 从高到低
排名模型供应商上下文Elo胜率源排名
1 Anthropic: Claude Opus 4.7anthropic/claude-opus-4.7 anthropic 1M 1331 65.2% 1
2 Anthropic: Claude Opus 4.7 (batch)anthropic/claude-opus-4.7:batch anthropic 1M 1331 65.2% 1
3 Anthropic: Claude Opus 4.8anthropic/claude-opus-4.8 anthropic 1M 1310 68.9% 2
4 Anthropic: Claude Opus 4.8 (batch)anthropic/claude-opus-4.8:batch anthropic 1M 1310 68.9% 2
5 Google: Gemini 3.5 Flashgoogle/gemini-3.5-flash google 1M 1242 57.8% 3
6 Google: Gemini 3.5 Flash (batch)google/gemini-3.5-flash:batch google 1M 1242 57.8% 3
7 Z.ai: GLM 5.1z-ai/glm-5.1 z-ai 204.8K 1240 53.3% 4
8 MoonshotAI: Kimi K2.6moonshotai/kimi-k2.6 moonshotai 262.1K 1186 45.5% 5
9 Z.ai: GLM 5V Turboz-ai/glm-5v-turbo z-ai 202.8K 1183 53.9% 6
10 OpenAI: GPT-5.5openai/gpt-5.5 openai 1.1M 1155 45.2% 7
11 OpenAI: GPT-5.5 (batch)openai/gpt-5.5:batch openai 1.1M 1155 45.2% 7
12 Google: Gemini 3.1 Pro Previewgoogle/gemini-3.1-pro-preview google 1M 1107 33.9% 8
13 Google: Gemini 3.1 Pro Preview (batch)google/gemini-3.1-pro-preview:batch google 1M 1107 33.9% 8
14 Google: Gemini 3 Flash Previewgoogle/gemini-3-flash-preview google 1M 1075 39.3% 9
15 Google: Gemini 3 Flash Preview (batch)google/gemini-3-flash-preview:batch google 1M 1075 39.3% 9
16 SpaceXAI: Grok 4.3x-ai/grok-4.3 x-ai 1M 1068 32.4% 10

Agent 场景 / Android 原生

Elo 从高到低
排名模型供应商上下文Elo胜率源排名
1 Anthropic: Claude Fable 5anthropic/claude-fable-5 anthropic 1M 1303 64.9% 1
2 Anthropic: Claude Fable 5 (batch)anthropic/claude-fable-5:batch anthropic 1M 1303 64.9% 1
3 SpaceXAI: Grok 4.5x-ai/grok-4.5 x-ai 500K 1272 68% 2
4 Z.ai: GLM 5V Turboz-ai/glm-5v-turbo z-ai 202.8K 1267 54.8% 3
5 Anthropic: Claude Opus 4.8anthropic/claude-opus-4.8 anthropic 1M 1261 58.8% 4
6 Anthropic: Claude Opus 4.8 (batch)anthropic/claude-opus-4.8:batch anthropic 1M 1261 58.8% 4
7 Anthropic: Claude Opus 4.7anthropic/claude-opus-4.7 anthropic 1M 1258 56% 5
8 Anthropic: Claude Opus 4.7 (batch)anthropic/claude-opus-4.7:batch anthropic 1M 1258 56% 5
9 Anthropic: Claude Sonnet 5anthropic/claude-sonnet-5 anthropic 1M 1254 56.7% 6
10 Anthropic: Claude Sonnet 5 (batch)anthropic/claude-sonnet-5:batch anthropic 1M 1254 56.7% 6
11 MoonshotAI: Kimi K3moonshotai/kimi-k3 moonshotai 1M 1250 56% 7
12 Meta: Muse Spark 1.2meta/muse-spark-1.2 meta 1M 1242 54.4% 8
13 Anthropic: Claude Opus 4.6anthropic/claude-opus-4.6 anthropic 1M 1239 66.5% 9
14 Anthropic: Claude Opus 4.6 (batch)anthropic/claude-opus-4.6:batch anthropic 1M 1239 66.5% 9
15 Anthropic: Claude Sonnet 4.6anthropic/claude-sonnet-4.6 anthropic 1M 1229 62.1% 10
16 Anthropic: Claude Sonnet 4.6 (batch)anthropic/claude-sonnet-4.6:batch anthropic 1M 1229 62.1% 10
17 Google: Gemini 3.6 Flashgoogle/gemini-3.6-flash google 1M 1226 53.8% 11
18 Google: Gemini 3.6 Flash (batch)google/gemini-3.6-flash:batch google 1M 1226 53.8% 11
19 Meta: Muse Spark 1.1meta/muse-spark-1.1 meta 1M 1218 49.8% 12
20 Google: Gemini 3.5 Flashgoogle/gemini-3.5-flash google 1M 1217 55.6% 13
21 Google: Gemini 3.5 Flash (batch)google/gemini-3.5-flash:batch google 1M 1217 55.6% 13
22 Z.ai: GLM 5.1z-ai/glm-5.1 z-ai 204.8K 1213 53.1% 14
23 Anthropic: Claude Opus 4.5anthropic/claude-opus-4.5 anthropic 200K 1209 65.5% 15
24 Anthropic: Claude Opus 4.5 (batch)anthropic/claude-opus-4.5:batch anthropic 200K 1209 65.5% 15
25 Z.ai: GLM 5.2z-ai/glm-5.2 z-ai 1M 1208 54.9% 16
26 Z.ai: GLM 5.2 (batch)z-ai/glm-5.2:batch z-ai 512K 1208 54.9% 16
27 OpenAI: GPT-5.5openai/gpt-5.5 openai 1.1M 1195 50.8% 17
28 OpenAI: GPT-5.5 (batch)openai/gpt-5.5:batch openai 1.1M 1195 50.8% 17
29 MoonshotAI: Kimi K2.6moonshotai/kimi-k2.6 moonshotai 262.1K 1195 51.5% 18
30 Qwen: Qwen3.7 Maxqwen/qwen3.7-max qwen 1M 1190 49.6% 19

Agent 场景 / godotgamedev

Elo 从高到低
排名模型供应商上下文Elo胜率源排名
1 Anthropic: Claude Fable 5anthropic/claude-fable-5 anthropic 1M 1345 70.3% 1
2 Anthropic: Claude Fable 5 (batch)anthropic/claude-fable-5:batch anthropic 1M 1345 70.3% 1
3 Anthropic: Claude Opus 4.7anthropic/claude-opus-4.7 anthropic 1M 1270 60.9% 2
4 Anthropic: Claude Opus 4.7 (batch)anthropic/claude-opus-4.7:batch anthropic 1M 1270 60.9% 2
5 SpaceXAI: Grok 4.5x-ai/grok-4.5 x-ai 500K 1270 62.1% 3
6 Anthropic: Claude Sonnet 5anthropic/claude-sonnet-5 anthropic 1M 1268 60.3% 4
7 Anthropic: Claude Sonnet 5 (batch)anthropic/claude-sonnet-5:batch anthropic 1M 1268 60.3% 4
8 Anthropic: Claude Opus 4.8anthropic/claude-opus-4.8 anthropic 1M 1254 58.9% 5
9 Anthropic: Claude Opus 4.8 (batch)anthropic/claude-opus-4.8:batch anthropic 1M 1254 58.9% 5
10 Google: Gemini 3.1 Pro Previewgoogle/gemini-3.1-pro-preview google 1M 1236 60% 6
11 Google: Gemini 3.1 Pro Preview (batch)google/gemini-3.1-pro-preview:batch google 1M 1236 60% 6
12 Anthropic: Claude Sonnet 4.6anthropic/claude-sonnet-4.6 anthropic 1M 1226 60.6% 7
13 Anthropic: Claude Sonnet 4.6 (batch)anthropic/claude-sonnet-4.6:batch anthropic 1M 1226 60.6% 7
14 Qwen: Qwen3.7 Maxqwen/qwen3.7-max qwen 1M 1224 55.4% 8
15 MoonshotAI: Kimi K2.5moonshotai/kimi-k2.5 moonshotai 262.1K 1219 59.8% 9
16 OpenAI: GPT-5.5openai/gpt-5.5 openai 1.1M 1212 52.4% 10
17 OpenAI: GPT-5.5 (batch)openai/gpt-5.5:batch openai 1.1M 1212 52.4% 10
18 MoonshotAI: Kimi K3moonshotai/kimi-k3 moonshotai 1M 1199 48.5% 11
19 MoonshotAI: Kimi K2.7 Codemoonshotai/kimi-k2.7-code moonshotai 262.1K 1184 49.2% 12
20 MoonshotAI: Kimi K2.7 Code (batch)moonshotai/kimi-k2.7-code:batch moonshotai 262.1K 1184 49.2% 12
21 Z.ai: GLM 4.6z-ai/glm-4.6 z-ai 204.8K 1180 53.2% 13
22 Google: Gemini 3 Flash Previewgoogle/gemini-3-flash-preview google 1M 1161 50.6% 14
23 Google: Gemini 3 Flash Preview (batch)google/gemini-3-flash-preview:batch google 1M 1161 50.6% 14
24 MoonshotAI: Kimi K2.6moonshotai/kimi-k2.6 moonshotai 262.1K 1160 47.1% 15
25 Z.ai: GLM 5z-ai/glm-5 z-ai 204.8K 1144 46.7% 16
26 Z.ai: GLM 5.2z-ai/glm-5.2 z-ai 1M 1142 40.1% 17
27 Z.ai: GLM 5.2 (batch)z-ai/glm-5.2:batch z-ai 512K 1142 40.1% 17
28 OpenAI: GPT-5.2openai/gpt-5.2 openai 400K 1142 48.1% 18
29 OpenAI: GPT-5.2 (batch)openai/gpt-5.2:batch openai 400K 1142 48.1% 18
30 OpenAI: GPT-5.2-Codexopenai/gpt-5.2-codex openai 400K 1142 47.8% 19

Agent 场景 / HTML 幻灯片

Elo 从高到低
排名模型供应商上下文Elo胜率源排名
1 Anthropic: Claude Fable 5anthropic/claude-fable-5 anthropic 1M 1264 58.9% 1
2 Anthropic: Claude Fable 5 (batch)anthropic/claude-fable-5:batch anthropic 1M 1264 58.9% 1
3 Anthropic: Claude Opus 4.7anthropic/claude-opus-4.7 anthropic 1M 1246 58.3% 2
4 Anthropic: Claude Opus 4.7 (batch)anthropic/claude-opus-4.7:batch anthropic 1M 1246 58.3% 2
5 Anthropic: Claude Opus 4.8anthropic/claude-opus-4.8 anthropic 1M 1235 56% 3
6 Anthropic: Claude Opus 4.8 (batch)anthropic/claude-opus-4.8:batch anthropic 1M 1235 56% 3
7 Anthropic: Claude Sonnet 5anthropic/claude-sonnet-5 anthropic 1M 1225 52.1% 4
8 Anthropic: Claude Sonnet 5 (batch)anthropic/claude-sonnet-5:batch anthropic 1M 1225 52.1% 4
9 SpaceXAI: Grok 4.5x-ai/grok-4.5 x-ai 500K 1219 52.8% 5
10 MoonshotAI: Kimi K2.7 Codemoonshotai/kimi-k2.7-code moonshotai 262.1K 1218 53.6% 6
11 MoonshotAI: Kimi K2.7 Code (batch)moonshotai/kimi-k2.7-code:batch moonshotai 262.1K 1218 53.6% 6
12 MoonshotAI: Kimi K2.6moonshotai/kimi-k2.6 moonshotai 262.1K 1217 53.4% 7
13 Meta: Muse Spark 1.1meta/muse-spark-1.1 meta 1M 1211 51.7% 8
14 Google: Gemini 3.1 Pro Previewgoogle/gemini-3.1-pro-preview google 1M 1194 50.8% 9
15 Google: Gemini 3.1 Pro Preview (batch)google/gemini-3.1-pro-preview:batch google 1M 1194 50.8% 9
16 Z.ai: GLM 5.2z-ai/glm-5.2 z-ai 1M 1193 49.3% 10
17 Z.ai: GLM 5.2 (batch)z-ai/glm-5.2:batch z-ai 512K 1193 49.3% 10
18 Z.ai: GLM 5.1z-ai/glm-5.1 z-ai 204.8K 1189 48.5% 11
19 Qwen: Qwen3.7 Maxqwen/qwen3.7-max qwen 1M 1186 44.5% 12
20 MiniMax: MiniMax M3minimax/minimax-m3 minimax 1M 1175 42.5% 13
21 MiniMax: MiniMax M3 (batch)minimax/minimax-m3:batch minimax 524.3K 1175 42.5% 13
22 SpaceXAI: Grok 4.20x-ai/grok-4.20 x-ai 2M 1174 45.7% 14
23 Google: Gemini 3.6 Flashgoogle/gemini-3.6-flash google 1M 1164 41% 15
24 Google: Gemini 3.6 Flash (batch)google/gemini-3.6-flash:batch google 1M 1164 41% 15
25 Google: Gemini 3.5 Flashgoogle/gemini-3.5-flash google 1M 1161 45.4% 16
26 Google: Gemini 3.5 Flash (batch)google/gemini-3.5-flash:batch google 1M 1161 45.4% 16
27 Z.ai: GLM 5z-ai/glm-5 z-ai 204.8K 1160 45.1% 17
28 Z.ai: GLM 5V Turboz-ai/glm-5v-turbo z-ai 202.8K 1144 43.6% 18
29 Meta: Muse Spark 1.2meta/muse-spark-1.2 meta 1M 1140 36.1% 19
30 OpenAI: GPT-5.5openai/gpt-5.5 openai 1.1M 1082 34.2% 20

Agent 场景 / pptxslides

Elo 从高到低
排名模型供应商上下文Elo胜率源排名
1 Anthropic: Claude Opus 4.7anthropic/claude-opus-4.7 anthropic 1M 1333 64.9% 1
2 Anthropic: Claude Opus 4.7 (batch)anthropic/claude-opus-4.7:batch anthropic 1M 1333 64.9% 1
3 Anthropic: Claude Opus 4.8anthropic/claude-opus-4.8 anthropic 1M 1306 67.9% 2
4 Anthropic: Claude Opus 4.8 (batch)anthropic/claude-opus-4.8:batch anthropic 1M 1306 67.9% 2
5 Google: Gemini 3.5 Flashgoogle/gemini-3.5-flash google 1M 1244 57.7% 3
6 Google: Gemini 3.5 Flash (batch)google/gemini-3.5-flash:batch google 1M 1244 57.7% 3
7 Z.ai: GLM 5.1z-ai/glm-5.1 z-ai 204.8K 1241 53.5% 4
8 MoonshotAI: Kimi K2.6moonshotai/kimi-k2.6 moonshotai 262.1K 1181 44.3% 5
9 Z.ai: GLM 5V Turboz-ai/glm-5v-turbo z-ai 202.8K 1164 52.1% 6
10 OpenAI: GPT-5.5openai/gpt-5.5 openai 1.1M 1157 45.3% 7
11 OpenAI: GPT-5.5 (batch)openai/gpt-5.5:batch openai 1.1M 1157 45.3% 7
12 Google: Gemini 3.1 Pro Previewgoogle/gemini-3.1-pro-preview google 1M 1110 34.1% 8
13 Google: Gemini 3.1 Pro Preview (batch)google/gemini-3.1-pro-preview:batch google 1M 1110 34.1% 8
14 SpaceXAI: Grok 4.3x-ai/grok-4.3 x-ai 1M 1071 32.5% 9

Agent 场景 / python-pptxslides

Elo 从高到低
排名模型供应商上下文Elo胜率源排名
1 MoonshotAI: Kimi K3moonshotai/kimi-k3 moonshotai 1M 1378 70.3% 1
2 Anthropic: Claude Opus 4.7anthropic/claude-opus-4.7 anthropic 1M 1336 63.1% 2
3 Anthropic: Claude Opus 4.7 (batch)anthropic/claude-opus-4.7:batch anthropic 1M 1336 63.1% 2
4 Anthropic: Claude Fable 5anthropic/claude-fable-5 anthropic 1M 1304 62% 3
5 Anthropic: Claude Fable 5 (batch)anthropic/claude-fable-5:batch anthropic 1M 1304 62% 3
6 Anthropic: Claude Opus 4.8anthropic/claude-opus-4.8 anthropic 1M 1298 65.9% 4
7 Anthropic: Claude Opus 4.8 (batch)anthropic/claude-opus-4.8:batch anthropic 1M 1298 65.9% 4
8 Z.ai: GLM 5.1z-ai/glm-5.1 z-ai 204.8K 1258 54.2% 5
9 Anthropic: Claude Sonnet 5anthropic/claude-sonnet-5 anthropic 1M 1250 53.8% 6
10 Anthropic: Claude Sonnet 5 (batch)anthropic/claude-sonnet-5:batch anthropic 1M 1250 53.8% 6
11 Google: Gemini 3.5 Flashgoogle/gemini-3.5-flash google 1M 1247 57.4% 7
12 Google: Gemini 3.5 Flash (batch)google/gemini-3.5-flash:batch google 1M 1247 57.4% 7
13 SpaceXAI: Grok 4.5x-ai/grok-4.5 x-ai 500K 1241 54.1% 8
14 MiniMax: MiniMax M3minimax/minimax-m3 minimax 1M 1227 49.2% 9
15 MiniMax: MiniMax M3 (batch)minimax/minimax-m3:batch minimax 524.3K 1227 49.2% 9
16 Qwen: Qwen3.7 Maxqwen/qwen3.7-max qwen 1M 1217 49% 10
17 Z.ai: GLM 5.2z-ai/glm-5.2 z-ai 1M 1197 47% 11
18 Z.ai: GLM 5.2 (batch)z-ai/glm-5.2:batch z-ai 512K 1197 47% 11
19 MoonshotAI: Kimi K2.6moonshotai/kimi-k2.6 moonshotai 262.1K 1180 42.1% 12
20 Meta: Muse Spark 1.1meta/muse-spark-1.1 meta 1M 1177 44.2% 13
21 Meta: Muse Spark 1.2meta/muse-spark-1.2 meta 1M 1168 41.4% 14
22 Z.ai: GLM 5V Turboz-ai/glm-5v-turbo z-ai 202.8K 1165 51.9% 15
23 MoonshotAI: Kimi K2.7 Codemoonshotai/kimi-k2.7-code moonshotai 262.1K 1161 43.5% 16
24 MoonshotAI: Kimi K2.7 Code (batch)moonshotai/kimi-k2.7-code:batch moonshotai 262.1K 1161 43.5% 16
25 OpenAI: GPT-5.5openai/gpt-5.5 openai 1.1M 1152 43.3% 17
26 OpenAI: GPT-5.5 (batch)openai/gpt-5.5:batch openai 1.1M 1152 43.3% 17
27 Google: Gemini 3.6 Flashgoogle/gemini-3.6-flash google 1M 1149 38.5% 18
28 Google: Gemini 3.6 Flash (batch)google/gemini-3.6-flash:batch google 1M 1149 38.5% 18
29 Google: Gemini 3.1 Pro Previewgoogle/gemini-3.1-pro-preview google 1M 1109 31.9% 19
30 Google: Gemini 3.1 Pro Preview (batch)google/gemini-3.1-pro-preview:batch google 1M 1109 31.9% 19

Agent 场景 / Web 应用

Elo 从高到低
排名模型供应商上下文Elo胜率源排名
1 MoonshotAI: Kimi K3moonshotai/kimi-k3 moonshotai 1M 1320 62.6% 1
2 Anthropic: Claude Opus 4.7anthropic/claude-opus-4.7 anthropic 1M 1295 62.1% 2
3 Anthropic: Claude Opus 4.7 (batch)anthropic/claude-opus-4.7:batch anthropic 1M 1295 62.1% 2
4 Anthropic: Claude Fable 5anthropic/claude-fable-5 anthropic 1M 1293 59.4% 3
5 Anthropic: Claude Fable 5 (batch)anthropic/claude-fable-5:batch anthropic 1M 1293 59.4% 3
6 Anthropic: Claude Sonnet 5anthropic/claude-sonnet-5 anthropic 1M 1290 58.3% 4
7 Anthropic: Claude Sonnet 5 (batch)anthropic/claude-sonnet-5:batch anthropic 1M 1290 58.3% 4
8 SpaceXAI: Grok 4.6x-ai/grok-4.6 x-ai 500K 1285 58.1% 5
9 Claude Opus 5anthropic/claude-opus-5 anthropic 1M 1279 56.3% 6
10 Claude Opus 5 (batch)anthropic/claude-opus-5:batch anthropic 1M 1279 56.3% 6
11 MoonshotAI: Kimi K2.6moonshotai/kimi-k2.6 moonshotai 262.1K 1268 59.3% 7
12 Anthropic: Claude Opus 4.8anthropic/claude-opus-4.8 anthropic 1M 1261 53.8% 8
13 Anthropic: Claude Opus 4.8 (batch)anthropic/claude-opus-4.8:batch anthropic 1M 1261 53.8% 8
14 Z.ai: GLM 5.2z-ai/glm-5.2 z-ai 1M 1258 56.9% 9
15 Z.ai: GLM 5.2 (batch)z-ai/glm-5.2:batch z-ai 512K 1258 56.9% 9
16 Meta: Muse Spark 1.2meta/muse-spark-1.2 meta 1M 1258 52.3% 10
17 SpaceXAI: Grok 4.5x-ai/grok-4.5 x-ai 500K 1250 54.2% 11
18 Qwen: Qwen3.7 Maxqwen/qwen3.7-max qwen 1M 1246 51.2% 12
19 Google: Gemini 3.6 Flashgoogle/gemini-3.6-flash google 1M 1241 50.4% 13
20 Google: Gemini 3.6 Flash (batch)google/gemini-3.6-flash:batch google 1M 1241 50.4% 13
21 Meta: Muse Spark 1.1meta/muse-spark-1.1 meta 1M 1239 50.4% 14
22 MiniMax: MiniMax M3minimax/minimax-m3 minimax 1M 1235 50.5% 15
23 MiniMax: MiniMax M3 (batch)minimax/minimax-m3:batch minimax 524.3K 1235 50.5% 15
24 Google: Gemini 3.5 Flashgoogle/gemini-3.5-flash google 1M 1233 52.6% 16
25 Google: Gemini 3.5 Flash (batch)google/gemini-3.5-flash:batch google 1M 1233 52.6% 16
26 Anthropic: Claude Opus 4.6anthropic/claude-opus-4.6 anthropic 1M 1230 56.2% 17
27 Anthropic: Claude Opus 4.6 (batch)anthropic/claude-opus-4.6:batch anthropic 1M 1230 56.2% 17
28 Anthropic: Claude Sonnet 4.6anthropic/claude-sonnet-4.6 anthropic 1M 1225 55.3% 18
29 Anthropic: Claude Sonnet 4.6 (batch)anthropic/claude-sonnet-4.6:batch anthropic 1M 1225 55.3% 18
30 Z.ai: GLM 5.1z-ai/glm-5.1 z-ai 204.8K 1214 53.1% 19

Agent 场景 / 全栈应用

Elo 从高到低
排名模型供应商上下文Elo胜率源排名
1 Anthropic: Claude Opus 4.7anthropic/claude-opus-4.7 anthropic 1M 1503 80.1% 1
2 Anthropic: Claude Opus 4.7 (batch)anthropic/claude-opus-4.7:batch anthropic 1M 1503 80.1% 1
3 Anthropic: Claude Fable 5anthropic/claude-fable-5 anthropic 1M 1286 61.1% 2
4 Anthropic: Claude Fable 5 (batch)anthropic/claude-fable-5:batch anthropic 1M 1286 61.1% 2
5 SpaceXAI: Grok 4.6x-ai/grok-4.6 x-ai 500K 1286 54.5% 3
6 SpaceXAI: Grok 4.5x-ai/grok-4.5 x-ai 500K 1280 64.1% 4
7 Anthropic: Claude Opus 4.8anthropic/claude-opus-4.8 anthropic 1M 1278 60.1% 5
8 Anthropic: Claude Opus 4.8 (batch)anthropic/claude-opus-4.8:batch anthropic 1M 1278 60.1% 5
9 Anthropic: Claude Sonnet 5anthropic/claude-sonnet-5 anthropic 1M 1271 57.7% 6
10 Anthropic: Claude Sonnet 5 (batch)anthropic/claude-sonnet-5:batch anthropic 1M 1271 57.7% 6
11 Z.ai: GLM 5.2z-ai/glm-5.2 z-ai 1M 1268 62.3% 7
12 Z.ai: GLM 5.2 (batch)z-ai/glm-5.2:batch z-ai 512K 1268 62.3% 7
13 Anthropic: Claude Opus 4.6anthropic/claude-opus-4.6 anthropic 1M 1252 64% 8
14 Anthropic: Claude Opus 4.6 (batch)anthropic/claude-opus-4.6:batch anthropic 1M 1252 64% 8
15 Anthropic: Claude Sonnet 4.6anthropic/claude-sonnet-4.6 anthropic 1M 1245 64.1% 9
16 Anthropic: Claude Sonnet 4.6 (batch)anthropic/claude-sonnet-4.6:batch anthropic 1M 1245 64.1% 9
17 Meta: Muse Spark 1.1meta/muse-spark-1.1 meta 1M 1245 50.9% 10
18 Meta: Muse Spark 1.2meta/muse-spark-1.2 meta 1M 1243 50% 11
19 Google: Gemini 3.5 Flashgoogle/gemini-3.5-flash google 1M 1228 56% 12
20 Google: Gemini 3.5 Flash (batch)google/gemini-3.5-flash:batch google 1M 1228 56% 12
21 MiniMax: MiniMax M3minimax/minimax-m3 minimax 1M 1221 50.6% 13
22 MiniMax: MiniMax M3 (batch)minimax/minimax-m3:batch minimax 524.3K 1221 50.6% 13
23 Qwen: Qwen3.7 Maxqwen/qwen3.7-max qwen 1M 1220 52.1% 14
24 MoonshotAI: Kimi K2.7 Codemoonshotai/kimi-k2.7-code moonshotai 262.1K 1208 53.5% 15
25 MoonshotAI: Kimi K2.7 Code (batch)moonshotai/kimi-k2.7-code:batch moonshotai 262.1K 1208 53.5% 15
26 Google: Gemini 3.6 Flashgoogle/gemini-3.6-flash google 1M 1202 47.1% 16
27 Google: Gemini 3.6 Flash (batch)google/gemini-3.6-flash:batch google 1M 1202 47.1% 16
28 Z.ai: GLM 5.1z-ai/glm-5.1 z-ai 204.8K 1200 55.4% 17
29 Anthropic: Claude Opus 4.5anthropic/claude-opus-4.5 anthropic 200K 1190 59.9% 18
30 Anthropic: Claude Opus 4.5 (batch)anthropic/claude-opus-4.5:batch anthropic 200K 1190 59.9% 18

Agent 场景 / 移动应用

Elo 从高到低
排名模型供应商上下文Elo胜率源排名
1 MoonshotAI: Kimi K3moonshotai/kimi-k3 moonshotai 1M 1301 60.5% 1
2 SpaceXAI: Grok 4.6x-ai/grok-4.6 x-ai 500K 1272 56.2% 2
3 Anthropic: Claude Opus 4.6anthropic/claude-opus-4.6 anthropic 1M 1268 63.3% 3
4 Anthropic: Claude Opus 4.6 (batch)anthropic/claude-opus-4.6:batch anthropic 1M 1268 63.3% 3
5 Anthropic: Claude Sonnet 4.6anthropic/claude-sonnet-4.6 anthropic 1M 1265 63.4% 4
6 Anthropic: Claude Sonnet 4.6 (batch)anthropic/claude-sonnet-4.6:batch anthropic 1M 1265 63.4% 4
7 SpaceXAI: Grok 4.5x-ai/grok-4.5 x-ai 500K 1259 56.7% 5
8 Anthropic: Claude Fable 5anthropic/claude-fable-5 anthropic 1M 1256 57% 6
9 Anthropic: Claude Fable 5 (batch)anthropic/claude-fable-5:batch anthropic 1M 1256 57% 6
10 Anthropic: Claude Opus 4.8anthropic/claude-opus-4.8 anthropic 1M 1248 56% 7
11 Anthropic: Claude Opus 4.8 (batch)anthropic/claude-opus-4.8:batch anthropic 1M 1248 56% 7
12 Google: Gemini 3.6 Flashgoogle/gemini-3.6-flash google 1M 1246 53.9% 8
13 Google: Gemini 3.6 Flash (batch)google/gemini-3.6-flash:batch google 1M 1246 53.9% 8
14 Anthropic: Claude Opus 4.5anthropic/claude-opus-4.5 anthropic 200K 1242 60.5% 9
15 Anthropic: Claude Opus 4.5 (batch)anthropic/claude-opus-4.5:batch anthropic 200K 1242 60.5% 9
16 Anthropic: Claude Sonnet 5anthropic/claude-sonnet-5 anthropic 1M 1235 52.5% 10
17 Anthropic: Claude Sonnet 5 (batch)anthropic/claude-sonnet-5:batch anthropic 1M 1235 52.5% 10
18 MiniMax: MiniMax M3minimax/minimax-m3 minimax 1M 1223 51% 11
19 MiniMax: MiniMax M3 (batch)minimax/minimax-m3:batch minimax 524.3K 1223 51% 11
20 Google: Gemini 3.5 Flashgoogle/gemini-3.5-flash google 1M 1221 53.9% 12
21 Google: Gemini 3.5 Flash (batch)google/gemini-3.5-flash:batch google 1M 1221 53.9% 12
22 MoonshotAI: Kimi K2.6moonshotai/kimi-k2.6 moonshotai 262.1K 1217 54.1% 13
23 Meta: Muse Spark 1.1meta/muse-spark-1.1 meta 1M 1216 48.9% 14
24 Z.ai: GLM 5.2z-ai/glm-5.2 z-ai 1M 1214 52.8% 15
25 Z.ai: GLM 5.2 (batch)z-ai/glm-5.2:batch z-ai 512K 1214 52.8% 15
26 Meta: Muse Spark 1.2meta/muse-spark-1.2 meta 1M 1213 47.7% 16
27 Z.ai: GLM 5.1z-ai/glm-5.1 z-ai 204.8K 1211 54% 17
28 Qwen: Qwen3.7 Maxqwen/qwen3.7-max qwen 1M 1206 49.4% 18
29 OpenAI: GPT-5.5openai/gpt-5.5 openai 1.1M 1205 51.7% 19
30 OpenAI: GPT-5.5 (batch)openai/gpt-5.5:batch openai 1.1M 1205 51.7% 19

模型场景 / 3D

Elo 从高到低
排名模型供应商上下文Elo胜率源排名
1 MoonshotAI: Kimi K3moonshotai/kimi-k3 moonshotai 1M 1453 69.2% 1
2 Claude Opus 5anthropic/claude-opus-5 anthropic 1M 1390 63.8% 2
3 Claude Opus 5 (batch)anthropic/claude-opus-5:batch anthropic 1M 1390 63.8% 2
4 Qwen: Qwen3.8 Maxqwen/qwen3.8-max qwen 1M 1388 59.6% 3
5 Anthropic: Claude Fable 5anthropic/claude-fable-5 anthropic 1M 1374 62.5% 4
6 Anthropic: Claude Fable 5 (batch)anthropic/claude-fable-5:batch anthropic 1M 1374 62.5% 4
7 SpaceXAI: Grok 4.6x-ai/grok-4.6 x-ai 500K 1368 58.6% 5
8 Z.ai: GLM 5.2z-ai/glm-5.2 z-ai 1M 1364 59.4% 6
9 Z.ai: GLM 5.2 (batch)z-ai/glm-5.2:batch z-ai 512K 1364 59.4% 6
10 Meta: Muse Spark 1.2meta/muse-spark-1.2 meta 1M 1356 62.6% 7
11 Z.ai: GLM 5.1z-ai/glm-5.1 z-ai 204.8K 1336 62.6% 8
12 Anthropic: Claude Opus 4.6anthropic/claude-opus-4.6 anthropic 1M 1330 62.4% 9
13 Anthropic: Claude Opus 4.6 (batch)anthropic/claude-opus-4.6:batch anthropic 1M 1330 62.4% 9
14 MoonshotAI: Kimi K2.6moonshotai/kimi-k2.6 moonshotai 262.1K 1329 59.9% 10
15 Google: Gemini 3.6 Flashgoogle/gemini-3.6-flash google 1M 1328 54.4% 11
16 Google: Gemini 3.6 Flash (batch)google/gemini-3.6-flash:batch google 1M 1328 54.4% 11
17 Qwen: Qwen3.7 Maxqwen/qwen3.7-max qwen 1M 1322 56.4% 12
18 SpaceXAI: Grok 4.5x-ai/grok-4.5 x-ai 500K 1312 50.3% 13
19 DeepSeek: DeepSeek V4 Prodeepseek/deepseek-v4-pro deepseek 1M 1310 57.4% 14
20 Anthropic: Claude Sonnet 5anthropic/claude-sonnet-5 anthropic 1M 1308 55.8% 15
21 Anthropic: Claude Sonnet 5 (batch)anthropic/claude-sonnet-5:batch anthropic 1M 1308 55.8% 15
22 Meta: Muse Spark 1.1meta/muse-spark-1.1 meta 1M 1308 53.4% 16
23 Z.ai: GLM 5 Turboz-ai/glm-5-turbo z-ai 202.8K 1306 57.9% 17
24 Nex AGI: Nex-N2-Pronex-agi/nex-n2-pro nex-agi 262.1K 1306 53.3% 18
25 Anthropic: Claude Opus 4.7anthropic/claude-opus-4.7 anthropic 1M 1305 56.8% 19
26 Anthropic: Claude Opus 4.7 (batch)anthropic/claude-opus-4.7:batch anthropic 1M 1305 56.8% 19
27 Xiaomi: MiMo-V2.5-Proxiaomi/mimo-v2.5-pro xiaomi 1.1M 1300 55.7% 20
28 MoonshotAI: Kimi K2.7 Codemoonshotai/kimi-k2.7-code moonshotai 262.1K 1298 52.2% 21
29 MoonshotAI: Kimi K2.7 Code (batch)moonshotai/kimi-k2.7-code:batch moonshotai 262.1K 1298 52.2% 21
30 Google: Gemini 3.5 Flashgoogle/gemini-3.5-flash google 1M 1296 57.8% 22

模型场景 / asciiart

Elo 从高到低
排名模型供应商上下文Elo胜率源排名
1 Anthropic: Claude Fable 5anthropic/claude-fable-5 anthropic 1M 1349 68.6% 1
2 Anthropic: Claude Fable 5 (batch)anthropic/claude-fable-5:batch anthropic 1M 1349 68.6% 1
3 Meta: Muse Spark 1.1meta/muse-spark-1.1 meta 1M 1333 62% 2
4 Anthropic: Claude Opus 4.7anthropic/claude-opus-4.7 anthropic 1M 1321 66.1% 3
5 Anthropic: Claude Opus 4.7 (batch)anthropic/claude-opus-4.7:batch anthropic 1M 1321 66.1% 3
6 Google: Gemini 3.1 Pro Previewgoogle/gemini-3.1-pro-preview google 1M 1309 63.6% 4
7 Google: Gemini 3.1 Pro Preview (batch)google/gemini-3.1-pro-preview:batch google 1M 1309 63.6% 4
8 Anthropic: Claude Opus 4.8anthropic/claude-opus-4.8 anthropic 1M 1306 62.9% 5
9 Anthropic: Claude Opus 4.8 (batch)anthropic/claude-opus-4.8:batch anthropic 1M 1306 62.9% 5
10 SpaceXAI: Grok 4.5x-ai/grok-4.5 x-ai 500K 1301 58.5% 6
11 Google: Gemini 3.5 Flashgoogle/gemini-3.5-flash google 1M 1295 59.5% 7
12 Google: Gemini 3.5 Flash (batch)google/gemini-3.5-flash:batch google 1M 1295 59.5% 7
13 Anthropic: Claude Opus 4.6anthropic/claude-opus-4.6 anthropic 1M 1291 63% 8
14 Anthropic: Claude Opus 4.6 (batch)anthropic/claude-opus-4.6:batch anthropic 1M 1291 63% 8
15 OpenAI: GPT-5.5openai/gpt-5.5 openai 1.1M 1286 60.7% 9
16 OpenAI: GPT-5.5 (batch)openai/gpt-5.5:batch openai 1.1M 1286 60.7% 9
17 Anthropic: Claude Sonnet 4.6anthropic/claude-sonnet-4.6 anthropic 1M 1271 60.1% 10
18 Anthropic: Claude Sonnet 4.6 (batch)anthropic/claude-sonnet-4.6:batch anthropic 1M 1271 60.1% 10
19 Z.ai: GLM 5.2z-ai/glm-5.2 z-ai 1M 1257 52% 11
20 Z.ai: GLM 5.2 (batch)z-ai/glm-5.2:batch z-ai 512K 1257 52% 11
21 Qwen: Qwen3.7 Maxqwen/qwen3.7-max qwen 1M 1251 53.5% 12
22 MoonshotAI: Kimi K2.7 Codemoonshotai/kimi-k2.7-code moonshotai 262.1K 1248 51.8% 13
23 MoonshotAI: Kimi K2.7 Code (batch)moonshotai/kimi-k2.7-code:batch moonshotai 262.1K 1248 51.8% 13
24 Anthropic: Claude Sonnet 5anthropic/claude-sonnet-5 anthropic 1M 1247 52.7% 14
25 Anthropic: Claude Sonnet 5 (batch)anthropic/claude-sonnet-5:batch anthropic 1M 1247 52.7% 14
26 Anthropic: Claude Sonnet 4.5anthropic/claude-sonnet-4.5 anthropic 1M 1241 56.1% 15
27 Anthropic: Claude Sonnet 4.5 (batch)anthropic/claude-sonnet-4.5:batch anthropic 1M 1241 56.1% 15
28 OpenAI: GPT-5.4openai/gpt-5.4 openai 1.1M 1238 55.5% 16
29 OpenAI: GPT-5.4 (batch)openai/gpt-5.4:batch openai 1.1M 1238 55.5% 16
30 Anthropic: Claude Opus 4.5anthropic/claude-opus-4.5 anthropic 200K 1227 54.7% 17

模型场景 / graphicdesign

Elo 从高到低
排名模型供应商上下文Elo胜率源排名
1 Google: Nano Banana 2 (Gemini 3.1 Flash Image Preview)google/gemini-3.1-flash-image-preview google 65.5K 1286 66.3% 2
2 Google: Nano Banana Pro (Gemini 3 Pro Image Preview)google/gemini-3-pro-image-preview google 65.5K 1275 65.8% 3
3 OpenAI: GPT-5 Imageopenai/gpt-5-image openai 400K 1197 48.9% 7
4 Google: Nano Banana (Gemini 2.5 Flash Image)google/gemini-2.5-flash-image google 32.8K 1192 56.9% 8
5 OpenAI: GPT-5 Image Miniopenai/gpt-5-image-mini openai 400K 1190 47.7% 9

模型场景 / image

Elo 从高到低
排名模型供应商上下文Elo胜率源排名
1 Google: Nano Banana 2 (Gemini 3.1 Flash Image Preview)google/gemini-3.1-flash-image-preview google 65.5K 1296 65.1% 2
2 Google: Nano Banana Pro (Gemini 3 Pro Image Preview)google/gemini-3-pro-image-preview google 65.5K 1263 62.1% 3
3 OpenAI: GPT-5 Imageopenai/gpt-5-image openai 400K 1211 53.9% 7
4 Google: Nano Banana (Gemini 2.5 Flash Image)google/gemini-2.5-flash-image google 32.8K 1206 55.6% 8
5 OpenAI: GPT-5 Image Miniopenai/gpt-5-image-mini openai 400K 1203 51% 9

模型场景 / imageediting

Elo 从高到低
排名模型供应商上下文Elo胜率源排名
1 Google: Nano Banana Pro (Gemini 3 Pro Image Preview)google/gemini-3-pro-image-preview google 65.5K 1270 65.4% 2

模型场景 / logo

Elo 从高到低
排名模型供应商上下文Elo胜率源排名
1 Google: Nano Banana 2 (Gemini 3.1 Flash Image Preview)google/gemini-3.1-flash-image-preview google 65.5K 1274 62.9% 2
2 Google: Nano Banana Pro (Gemini 3 Pro Image Preview)google/gemini-3-pro-image-preview google 65.5K 1252 61% 4
3 OpenAI: GPT-5 Image Miniopenai/gpt-5-image-mini openai 400K 1220 51.4% 6
4 OpenAI: GPT-5 Imageopenai/gpt-5-image openai 400K 1214 52.7% 7
5 Google: Nano Banana (Gemini 2.5 Flash Image)google/gemini-2.5-flash-image google 32.8K 1181 51.4% 9

模型场景 / svg

Elo 从高到低
排名模型供应商上下文Elo胜率源排名
1 Claude Opus 5anthropic/claude-opus-5 anthropic 1M 1368 60.7% 1
2 Claude Opus 5 (batch)anthropic/claude-opus-5:batch anthropic 1M 1368 60.7% 1
3 MoonshotAI: Kimi K3moonshotai/kimi-k3 moonshotai 1M 1342 64.6% 2
4 Anthropic: Claude Fable 5anthropic/claude-fable-5 anthropic 1M 1340 64.5% 3
5 Anthropic: Claude Fable 5 (batch)anthropic/claude-fable-5:batch anthropic 1M 1340 64.5% 3
6 Google: Gemini 3.1 Pro Previewgoogle/gemini-3.1-pro-preview google 1M 1321 68% 4
7 Google: Gemini 3.1 Pro Preview (batch)google/gemini-3.1-pro-preview:batch google 1M 1321 68% 4
8 Google: Gemini 3.5 Flashgoogle/gemini-3.5-flash google 1M 1284 60.3% 5
9 Google: Gemini 3.5 Flash (batch)google/gemini-3.5-flash:batch google 1M 1284 60.3% 5
10 OpenAI: GPT-5.5openai/gpt-5.5 openai 1.1M 1267 57.5% 6
11 OpenAI: GPT-5.5 (batch)openai/gpt-5.5:batch openai 1.1M 1267 57.5% 6
12 Anthropic: Claude Opus 4.6anthropic/claude-opus-4.6 anthropic 1M 1264 60.6% 7
13 Anthropic: Claude Opus 4.6 (batch)anthropic/claude-opus-4.6:batch anthropic 1M 1264 60.6% 7
14 Anthropic: Claude Opus 4.7anthropic/claude-opus-4.7 anthropic 1M 1260 58.9% 8
15 Anthropic: Claude Opus 4.7 (batch)anthropic/claude-opus-4.7:batch anthropic 1M 1260 58.9% 8
16 Meta: Muse Spark 1.1meta/muse-spark-1.1 meta 1M 1258 48.2% 9
17 SpaceXAI: Grok 4.5x-ai/grok-4.5 x-ai 500K 1255 50.2% 10
18 Qwen: Qwen3.7 Maxqwen/qwen3.7-max qwen 1M 1254 59.7% 11
19 Z.ai: GLM 5.1z-ai/glm-5.1 z-ai 204.8K 1253 59% 12
20 Z.ai: GLM 5.2z-ai/glm-5.2 z-ai 1M 1253 55.7% 13
21 Z.ai: GLM 5.2 (batch)z-ai/glm-5.2:batch z-ai 512K 1253 55.7% 13
22 Z.ai: GLM 5 Turboz-ai/glm-5-turbo z-ai 202.8K 1240 56.6% 14
23 Anthropic: Claude Sonnet 4.6anthropic/claude-sonnet-4.6 anthropic 1M 1233 59% 15
24 Anthropic: Claude Sonnet 4.6 (batch)anthropic/claude-sonnet-4.6:batch anthropic 1M 1233 59% 15
25 Nex AGI: Nex-N2-Pronex-agi/nex-n2-pro nex-agi 262.1K 1231 51.2% 16
26 OpenAI: GPT-5.4openai/gpt-5.4 openai 1.1M 1230 57.9% 17
27 OpenAI: GPT-5.4 (batch)openai/gpt-5.4:batch openai 1.1M 1230 57.9% 17
28 Anthropic: Claude Sonnet 5anthropic/claude-sonnet-5 anthropic 1M 1226 53.3% 18
29 Anthropic: Claude Sonnet 5 (batch)anthropic/claude-sonnet-5:batch anthropic 1M 1226 53.3% 18
30 OpenAI: GPT-5openai/gpt-5 openai 400K 1218 62.3% 19

模型场景 / UI 组件

Elo 从高到低
排名模型供应商上下文Elo胜率源排名
1 MoonshotAI: Kimi K3moonshotai/kimi-k3 moonshotai 1M 1386 63.3% 1
2 SpaceXAI: Grok 4.6x-ai/grok-4.6 x-ai 500K 1377 62.3% 2
3 Claude Opus 5anthropic/claude-opus-5 anthropic 1M 1369 59.5% 3
4 Claude Opus 5 (batch)anthropic/claude-opus-5:batch anthropic 1M 1369 59.5% 3
5 Qwen: Qwen3.8 Maxqwen/qwen3.8-max qwen 1M 1356 59.8% 4
6 Meta: Muse Spark 1.2meta/muse-spark-1.2 meta 1M 1344 58.3% 5
7 Anthropic: Claude Fable 5anthropic/claude-fable-5 anthropic 1M 1336 56.1% 6
8 Anthropic: Claude Fable 5 (batch)anthropic/claude-fable-5:batch anthropic 1M 1336 56.1% 6
9 Google: Gemini 3.6 Flashgoogle/gemini-3.6-flash google 1M 1334 55% 7
10 Google: Gemini 3.6 Flash (batch)google/gemini-3.6-flash:batch google 1M 1334 55% 7
11 Z.ai: GLM 5.2z-ai/glm-5.2 z-ai 1M 1334 56.8% 8
12 Z.ai: GLM 5.2 (batch)z-ai/glm-5.2:batch z-ai 512K 1334 56.8% 8
13 Anthropic: Claude Opus 4.7anthropic/claude-opus-4.7 anthropic 1M 1332 59.3% 9
14 Anthropic: Claude Opus 4.7 (batch)anthropic/claude-opus-4.7:batch anthropic 1M 1332 59.3% 9
15 Meta: Muse Spark 1.1meta/muse-spark-1.1 meta 1M 1323 53.5% 10
16 Anthropic: Claude Opus 4.6anthropic/claude-opus-4.6 anthropic 1M 1313 59.1% 11
17 Anthropic: Claude Opus 4.6 (batch)anthropic/claude-opus-4.6:batch anthropic 1M 1313 59.1% 11
18 SpaceXAI: Grok 4.5x-ai/grok-4.5 x-ai 500K 1313 52.3% 12
19 Anthropic: Claude Sonnet 5anthropic/claude-sonnet-5 anthropic 1M 1308 54.5% 13
20 Anthropic: Claude Sonnet 5 (batch)anthropic/claude-sonnet-5:batch anthropic 1M 1308 54.5% 13
21 Google: Gemini 3.1 Pro Previewgoogle/gemini-3.1-pro-preview google 1M 1306 69.3% 14
22 Google: Gemini 3.1 Pro Preview (batch)google/gemini-3.1-pro-preview:batch google 1M 1306 69.3% 14
23 Google: Gemini 3.5 Flashgoogle/gemini-3.5-flash google 1M 1303 56% 15
24 Google: Gemini 3.5 Flash (batch)google/gemini-3.5-flash:batch google 1M 1303 56% 15
25 Z.ai: GLM 5.1z-ai/glm-5.1 z-ai 204.8K 1302 54.3% 16
26 Qwen: Qwen3.7 Maxqwen/qwen3.7-max qwen 1M 1302 53.9% 17
27 Anthropic: Claude Sonnet 4.6anthropic/claude-sonnet-4.6 anthropic 1M 1299 57.3% 18
28 Anthropic: Claude Sonnet 4.6 (batch)anthropic/claude-sonnet-4.6:batch anthropic 1M 1299 57.3% 18
29 MoonshotAI: Kimi K2.6moonshotai/kimi-k2.6 moonshotai 262.1K 1296 56% 19
30 Xiaomi: MiMo-V2.5-Proxiaomi/mimo-v2.5-pro xiaomi 1.1M 1293 55.6% 20

模型场景 / 代码分类

Elo 从高到低
排名模型供应商上下文Elo胜率源排名
1 MoonshotAI: Kimi K3moonshotai/kimi-k3 moonshotai 1M 1404 65.9% 1
2 Claude Opus 5anthropic/claude-opus-5 anthropic 1M 1348 59.5% 2
3 Claude Opus 5 (batch)anthropic/claude-opus-5:batch anthropic 1M 1348 59.5% 2
4 Meta: Muse Spark 1.2meta/muse-spark-1.2 meta 1M 1341 59.4% 3
5 Anthropic: Claude Fable 5anthropic/claude-fable-5 anthropic 1M 1334 59.1% 4
6 Anthropic: Claude Fable 5 (batch)anthropic/claude-fable-5:batch anthropic 1M 1334 59.1% 4
7 SpaceXAI: Grok 4.6x-ai/grok-4.6 x-ai 500K 1333 56.5% 5
8 Z.ai: GLM 5.2z-ai/glm-5.2 z-ai 1M 1330 58.4% 6
9 Z.ai: GLM 5.2 (batch)z-ai/glm-5.2:batch z-ai 512K 1330 58.4% 6
10 Qwen: Qwen3.8 Maxqwen/qwen3.8-max qwen 1M 1322 56% 7
11 Google: Gemini 3.6 Flashgoogle/gemini-3.6-flash google 1M 1311 55.5% 8
12 Google: Gemini 3.6 Flash (batch)google/gemini-3.6-flash:batch google 1M 1311 55.5% 8
13 Anthropic: Claude Opus 4.7anthropic/claude-opus-4.7 anthropic 1M 1310 58.2% 9
14 Anthropic: Claude Opus 4.7 (batch)anthropic/claude-opus-4.7:batch anthropic 1M 1310 58.2% 9
15 Anthropic: Claude Opus 4.6anthropic/claude-opus-4.6 anthropic 1M 1309 61% 10
16 Anthropic: Claude Opus 4.6 (batch)anthropic/claude-opus-4.6:batch anthropic 1M 1309 61% 10
17 SpaceXAI: Grok 4.5x-ai/grok-4.5 x-ai 500K 1300 52.5% 11
18 Meta: Muse Spark 1.1meta/muse-spark-1.1 meta 1M 1298 53.5% 12
19 Anthropic: Claude Sonnet 4.6anthropic/claude-sonnet-4.6 anthropic 1M 1296 59.1% 13
20 Anthropic: Claude Sonnet 4.6 (batch)anthropic/claude-sonnet-4.6:batch anthropic 1M 1296 59.1% 13
21 MoonshotAI: Kimi K2.6moonshotai/kimi-k2.6 moonshotai 262.1K 1296 56.4% 14
22 Qwen: Qwen3.7 Maxqwen/qwen3.7-max qwen 1M 1296 55.9% 15
23 Anthropic: Claude Sonnet 5anthropic/claude-sonnet-5 anthropic 1M 1295 54.5% 16
24 Anthropic: Claude Sonnet 5 (batch)anthropic/claude-sonnet-5:batch anthropic 1M 1295 54.5% 16
25 Xiaomi: MiMo-V2.5-Proxiaomi/mimo-v2.5-pro xiaomi 1.1M 1292 54.6% 17
26 Z.ai: GLM 5.1z-ai/glm-5.1 z-ai 204.8K 1291 55.4% 18
27 Z.ai: GLM 5 Turboz-ai/glm-5-turbo z-ai 202.8K 1286 55.6% 19
28 Qwen: Qwen3.7 Plusqwen/qwen3.7-plus qwen 1M 1283 50.4% 20
29 Google: Gemini 3.5 Flashgoogle/gemini-3.5-flash google 1M 1282 55.8% 21
30 Google: Gemini 3.5 Flash (batch)google/gemini-3.5-flash:batch google 1M 1282 55.8% 21

模型场景 / 数据可视化

Elo 从高到低
排名模型供应商上下文Elo胜率源排名
1 MoonshotAI: Kimi K3moonshotai/kimi-k3 moonshotai 1M 1375 63.5% 1
2 Claude Opus 5anthropic/claude-opus-5 anthropic 1M 1374 61.4% 2
3 Claude Opus 5 (batch)anthropic/claude-opus-5:batch anthropic 1M 1374 61.4% 2
4 Z.ai: GLM 5.1z-ai/glm-5.1 z-ai 204.8K 1366 67% 3
5 Meta: Muse Spark 1.2meta/muse-spark-1.2 meta 1M 1360 62.8% 4
6 Anthropic: Claude Fable 5anthropic/claude-fable-5 anthropic 1M 1342 58.4% 5
7 Anthropic: Claude Fable 5 (batch)anthropic/claude-fable-5:batch anthropic 1M 1342 58.4% 5
8 SpaceXAI: Grok 4.6x-ai/grok-4.6 x-ai 500K 1341 57.1% 6
9 Google: Gemini 3.6 Flashgoogle/gemini-3.6-flash google 1M 1331 55.8% 7
10 Google: Gemini 3.6 Flash (batch)google/gemini-3.6-flash:batch google 1M 1331 55.8% 7
11 Z.ai: GLM 5.2z-ai/glm-5.2 z-ai 1M 1330 56.1% 8
12 Z.ai: GLM 5.2 (batch)z-ai/glm-5.2:batch z-ai 512K 1330 56.1% 8
13 Anthropic: Claude Opus 4.7anthropic/claude-opus-4.7 anthropic 1M 1309 56.8% 9
14 Anthropic: Claude Opus 4.7 (batch)anthropic/claude-opus-4.7:batch anthropic 1M 1309 56.8% 9
15 Anthropic: Claude Opus 4.6anthropic/claude-opus-4.6 anthropic 1M 1306 58.2% 10
16 Anthropic: Claude Opus 4.6 (batch)anthropic/claude-opus-4.6:batch anthropic 1M 1306 58.2% 10
17 Meta: Muse Spark 1.1meta/muse-spark-1.1 meta 1M 1305 52.4% 11
18 Anthropic: Claude Sonnet 4.6anthropic/claude-sonnet-4.6 anthropic 1M 1304 57.5% 12
19 Anthropic: Claude Sonnet 4.6 (batch)anthropic/claude-sonnet-4.6:batch anthropic 1M 1304 57.5% 12
20 Qwen: Qwen3.7 Maxqwen/qwen3.7-max qwen 1M 1304 54.8% 13
21 SpaceXAI: Grok 4.5x-ai/grok-4.5 x-ai 500K 1300 51.9% 14
22 Xiaomi: MiMo-V2.5-Proxiaomi/mimo-v2.5-pro xiaomi 1.1M 1289 51.7% 15
23 Z.ai: GLM 5 Turboz-ai/glm-5-turbo z-ai 202.8K 1287 57.2% 16
24 OpenAI: GPT-5.5openai/gpt-5.5 openai 1.1M 1281 56.8% 17
25 OpenAI: GPT-5.5 (batch)openai/gpt-5.5:batch openai 1.1M 1281 56.8% 17
26 MoonshotAI: Kimi K2.6moonshotai/kimi-k2.6 moonshotai 262.1K 1279 55.8% 18
27 Qwen: Qwen3.8 Maxqwen/qwen3.8-max qwen 1M 1279 52.8% 19
28 Xiaomi: MiMo-V2.5xiaomi/mimo-v2.5 xiaomi 1.1M 1275 55.2% 20
29 Qwen: Qwen3.7 Plusqwen/qwen3.7-plus qwen 1M 1273 51% 21
30 Anthropic: Claude Opus 4.5anthropic/claude-opus-4.5 anthropic 200K 1265 58.5% 22

模型场景 / 游戏开发

Elo 从高到低
排名模型供应商上下文Elo胜率源排名
1 MoonshotAI: Kimi K3moonshotai/kimi-k3 moonshotai 1M 1430 64% 1
2 Claude Opus 5anthropic/claude-opus-5 anthropic 1M 1395 63.1% 2
3 Claude Opus 5 (batch)anthropic/claude-opus-5:batch anthropic 1M 1395 63.1% 2
4 Anthropic: Claude Fable 5anthropic/claude-fable-5 anthropic 1M 1387 62.3% 3
5 Anthropic: Claude Fable 5 (batch)anthropic/claude-fable-5:batch anthropic 1M 1387 62.3% 3
6 Meta: Muse Spark 1.2meta/muse-spark-1.2 meta 1M 1361 60.5% 4
7 SpaceXAI: Grok 4.6x-ai/grok-4.6 x-ai 500K 1352 56.7% 5
8 Anthropic: Claude Sonnet 5anthropic/claude-sonnet-5 anthropic 1M 1348 55.3% 6
9 Anthropic: Claude Sonnet 5 (batch)anthropic/claude-sonnet-5:batch anthropic 1M 1348 55.3% 6
10 OpenAI: GPT-5.5openai/gpt-5.5 openai 1.1M 1341 59.6% 7
11 OpenAI: GPT-5.5 (batch)openai/gpt-5.5:batch openai 1.1M 1341 59.6% 7
12 Z.ai: GLM 5.2z-ai/glm-5.2 z-ai 1M 1338 56.8% 8
13 Z.ai: GLM 5.2 (batch)z-ai/glm-5.2:batch z-ai 512K 1338 56.8% 8
14 Qwen: Qwen3.8 Maxqwen/qwen3.8-max qwen 1M 1332 56.4% 9
15 Anthropic: Claude Opus 4.7anthropic/claude-opus-4.7 anthropic 1M 1331 59% 10
16 Anthropic: Claude Opus 4.7 (batch)anthropic/claude-opus-4.7:batch anthropic 1M 1331 59% 10
17 Meta: Muse Spark 1.1meta/muse-spark-1.1 meta 1M 1326 52.3% 11
18 Anthropic: Claude Opus 4.6anthropic/claude-opus-4.6 anthropic 1M 1319 59.3% 12
19 Anthropic: Claude Opus 4.6 (batch)anthropic/claude-opus-4.6:batch anthropic 1M 1319 59.3% 12
20 Google: Gemini 3.5 Flashgoogle/gemini-3.5-flash google 1M 1314 55.8% 13
21 Google: Gemini 3.5 Flash (batch)google/gemini-3.5-flash:batch google 1M 1314 55.8% 13
22 Qwen: Qwen3.7 Maxqwen/qwen3.7-max qwen 1M 1314 56% 14
23 SpaceXAI: Grok 4.5x-ai/grok-4.5 x-ai 500K 1312 50.8% 15
24 Xiaomi: MiMo-V2.5-Proxiaomi/mimo-v2.5-pro xiaomi 1.1M 1309 55.3% 16
25 Qwen: Qwen3.7 Plusqwen/qwen3.7-plus qwen 1M 1306 51.9% 17
26 Z.ai: GLM 5.1z-ai/glm-5.1 z-ai 204.8K 1305 56.5% 18
27 Anthropic: Claude Sonnet 4.6anthropic/claude-sonnet-4.6 anthropic 1M 1303 58.8% 19
28 Anthropic: Claude Sonnet 4.6 (batch)anthropic/claude-sonnet-4.6:batch anthropic 1M 1303 58.8% 19
29 Google: Gemini 3.6 Flashgoogle/gemini-3.6-flash google 1M 1300 53.6% 20
30 Google: Gemini 3.6 Flash (batch)google/gemini-3.6-flash:batch google 1M 1300 53.6% 20

模型场景 / 网站

Elo 从高到低
排名模型供应商上下文Elo胜率源排名
1 MoonshotAI: Kimi K3moonshotai/kimi-k3 moonshotai 1M 1369 63.1% 1
2 Meta: Muse Spark 1.2meta/muse-spark-1.2 meta 1M 1329 57.9% 2
3 Claude Opus 5anthropic/claude-opus-5 anthropic 1M 1328 58.3% 3
4 Claude Opus 5 (batch)anthropic/claude-opus-5:batch anthropic 1M 1328 58.3% 3
5 Z.ai: GLM 5.2z-ai/glm-5.2 z-ai 1M 1321 59% 4
6 Z.ai: GLM 5.2 (batch)z-ai/glm-5.2:batch z-ai 512K 1321 59% 4
7 SpaceXAI: Grok 4.6x-ai/grok-4.6 x-ai 500K 1318 55.3% 5
8 Anthropic: Claude Fable 5anthropic/claude-fable-5 anthropic 1M 1314 58.8% 6
9 Anthropic: Claude Fable 5 (batch)anthropic/claude-fable-5:batch anthropic 1M 1314 58.8% 6
10 Google: Gemini 3.6 Flashgoogle/gemini-3.6-flash google 1M 1314 57.6% 7
11 Google: Gemini 3.6 Flash (batch)google/gemini-3.6-flash:batch google 1M 1314 57.6% 7
12 Anthropic: Claude Opus 4.7anthropic/claude-opus-4.7 anthropic 1M 1306 58.6% 8
13 Anthropic: Claude Opus 4.7 (batch)anthropic/claude-opus-4.7:batch anthropic 1M 1306 58.6% 8
14 Anthropic: Claude Opus 4.6anthropic/claude-opus-4.6 anthropic 1M 1303 61.1% 9
15 Anthropic: Claude Opus 4.6 (batch)anthropic/claude-opus-4.6:batch anthropic 1M 1303 61.1% 9
16 SpaceXAI: Grok 4.5x-ai/grok-4.5 x-ai 500K 1303 55.7% 10
17 Qwen: Qwen3.8 Maxqwen/qwen3.8-max qwen 1M 1301 54.7% 11
18 Anthropic: Claude Sonnet 4.6anthropic/claude-sonnet-4.6 anthropic 1M 1296 59.8% 12
19 Anthropic: Claude Sonnet 4.6 (batch)anthropic/claude-sonnet-4.6:batch anthropic 1M 1296 59.8% 12
20 Z.ai: GLM 5.1z-ai/glm-5.1 z-ai 204.8K 1290 55.1% 13
21 Anthropic: Claude Sonnet 5anthropic/claude-sonnet-5 anthropic 1M 1286 54.6% 14
22 Anthropic: Claude Sonnet 5 (batch)anthropic/claude-sonnet-5:batch anthropic 1M 1286 54.6% 14
23 Xiaomi: MiMo-V2.5-Proxiaomi/mimo-v2.5-pro xiaomi 1.1M 1286 54.5% 15
24 Qwen: Qwen3.7 Maxqwen/qwen3.7-max qwen 1M 1285 56.1% 16
25 MoonshotAI: Kimi K2.6moonshotai/kimi-k2.6 moonshotai 262.1K 1284 55% 17
26 MoonshotAI: Kimi K2.7 Codemoonshotai/kimi-k2.7-code moonshotai 262.1K 1284 53.9% 18
27 MoonshotAI: Kimi K2.7 Code (batch)moonshotai/kimi-k2.7-code:batch moonshotai 262.1K 1284 53.9% 18
28 Meta: Muse Spark 1.1meta/muse-spark-1.1 meta 1M 1283 53.2% 19
29 Qwen: Qwen3.7 Plusqwen/qwen3.7-plus qwen 1M 1282 51.4% 20
30 Z.ai: GLM 5 Turboz-ai/glm-5-turbo z-ai 202.8K 1280 54.9% 21