模型榜单

这是一个基于 Chatbot Arena (lmarena.ai) 数据的排行榜,通过自动化流程生成。

数据更新时间: 2026-02-21 08:09:32 UTC / 2026-02-21 16:09:32 CST (北京时间)

排行榜

Rank
Rank Spread
模型
分数
票数

1

13

Anthropicclaude-opus-4-6Anthropic · Proprietary

1505±8

6,212

2

13

Anthropicclaude-opus-4-6-thinkingAnthropic · Proprietary

1504±8

5,350

3

13

gemini-3.1-pro-previewGoogle · Proprietary

1500±9Preliminary

4,072

4

45

gemini-3-proGoogle · Proprietary

1486±4

37,088

5

412

Bytedancedola-seed-2.0-previewBytedance · Proprietary

1475±10Preliminary

3,379

6

59

gemini-3-flashGoogle · Proprietary

1474±5

27,875

7

59

grok-4.1-thinkingxAI · Proprietary

1474±4

36,233

8

510

Anthropicclaude-opus-4-5-20251101-thinking-32kAnthropic · Proprietary

1471±4

29,183

9

512

Anthropicclaude-opus-4-5-20251101Anthropic · Proprietary

1467±4

34,111

10

816

grok-4.1xAI · Proprietary

1463±4

40,453

11

816

gemini-3-flash (thinking-minimal)Google · Proprietary

1462±5

19,226

12

1019

gpt-5.1-highOpenAI · Proprietary

1458±4

33,100

13

725

Anthropicclaude-sonnet-4-6Anthropic · Proprietary

1457±10

3,255

14

1025

glm-5Z.ai · MIT

1455±8

5,319

15

1025

ernie-5.0-0110Baidu · Proprietary

1453±6

12,706

16

1225

Anthropicclaude-sonnet-4-5-20250929Anthropic · Proprietary

1450±4

45,424

17

1030

qwen3.5-397b-a17bAlibaba · Apache 2.0

1450±10

3,602

18

1225

Anthropicclaude-sonnet-4-5-20250929-thinking-32kAnthropic · Proprietary

1450±4

47,652

19

1325

gemini-2.5-proGoogle · Proprietary

1449±3

96,334

20

1229

MoonshotAIkimi-k2.5-thinkingMoonshot · Modified MIT

1449±6

9,823

21

1229

ernie-5.0-preview-1203Baidu · Proprietary

1449±7

9,742

22

1327

Anthropicclaude-opus-4-1-20250805-thinking-16kAnthropic · Proprietary

1449±4

49,802

23

1329

Anthropicclaude-opus-4-1-20250805Anthropic · Proprietary

1446±3

76,656

24

1332

gpt-4.5-preview-2025-02-27OpenAI · Proprietary

1444±6

14,549

25

1930

chatgpt-4o-latest-20250326OpenAI · Proprietary

1442±3

83,247

26

1334

glm-4.7Z.ai · MIT

1441±6

11,968

27

1934

gpt-5.2-highOpenAI · Proprietary

1441±6

17,912

28

2035

gpt-5.2OpenAI · Proprietary

1438±6

14,700

29

2335

gpt-5.1OpenAI · Proprietary

1438±4

35,392

30

2040

MoonshotAIkimi-k2.5-instantMoonshot · Modified MIT

1437±8

5,857

31

2537

gpt-5-highOpenAI · Proprietary

1434±5

32,547

32

2539

qwen3-max-previewAlibaba · Proprietary

1434±5

27,755

33

2641

o3-2025-04-16OpenAI · Proprietary

1432±4

61,262

34

2643

grok-4-1-fast-reasoningxAI · Proprietary

1431±4

29,899

35

2844

MoonshotAIkimi-k2-thinking-turboMoonshot · Modified MIT

1430±4

34,970

36

3052

gpt-5-chatOpenAI · Proprietary

1426±4

31,744

37

3354

glm-4.6Z.ai · MIT

1425±4

35,233

38

3055

qwen3-max-2025-09-23Alibaba · Proprietary

1425±6

9,200

39

3454

Anthropicclaude-opus-4-20250514-thinking-16kAnthropic · Proprietary

1424±4

37,920

40

3157

deepseek-v3.2-expDeepSeek · MIT

1423±6

11,721

41

3057

deepseek-v3.2-exp-thinkingDeepSeek · MIT

1423±7

8,979

42

3654

qwen3-235b-a22b-instruct-2507Alibaba · Apache 2.0

1422±3

70,609

43

3259

grok-4-fast-chatxAI · Proprietary

1422±8

6,982

44

3657

deepseek-v3.2-thinkingDeepSeek · MIT

1420±5

24,580

45

3659

deepseek-v3.2DeepSeek · MIT

1419±5

29,555

46

3662

deepseek-r1-0528DeepSeek · MIT

1419±6

19,273

47

3462

ernie-5.0-preview-1022Baidu · Proprietary

1419±9

4,593

48

3662

deepseek-v3.1DeepSeek · MIT

1418±6

15,260

49

3662

MoonshotAIkimi-k2-0905-previewMoonshot · Modified MIT

1417±6

11,955

50

3662

deepseek-v3.1-thinkingDeepSeek · MIT

1417±7

11,959

51

3762

MoonshotAIkimi-k2-0711-previewMoonshot · Modified MIT

1417±5

28,625

52

3566

deepseek-v3.1-terminusDeepSeek · MIT

1416±10

3,756

53

3670

deepseek-v3.1-terminus-thinkingDeepSeek · MIT

1416±10

3,547

54

3766

qwen3-vl-235b-a22b-instructAlibaba · Apache 2.0

1415±6

11,649

55

4062

mistral-large-3Mistral · Apache 2.0

1414±5

25,814

56

4162

gpt-4.1-2025-04-14OpenAI · Proprietary

1413±4

52,108

57

4163

Anthropicclaude-opus-4-20250514Anthropic · Proprietary

1413±4

45,501

58

4466

grok-3-preview-02-24xAI · Proprietary

1411±4

33,960

59

4666

mistral-medium-2508Mistral · Proprietary

1411±3

64,525

60

4666

gemini-2.5-flashGoogle · Proprietary

1411±3

95,586

61

4471

glm-4.5Z.ai · MIT

1410±5

24,745

62

4670

grok-4-0709xAI · Proprietary

1410±4

41,971

63

5576

Anthropicclaude-haiku-4-5-20251001Anthropic · Proprietary

1405±4

46,152

64

5576

gemini-2.5-flash-preview-09-2025Google · Proprietary

1404±4

32,637

65

5576

grok-4-fast-reasoningxAI · Proprietary

1404±5

18,530

66

6077

o1-2024-12-17OpenAI · Proprietary

1402±4

27,822

67

6077

qwen3-235b-a22b-no-thinkingAlibaba · Apache 2.0

1402±4

39,501

68

6079

qwen3-next-80b-a3b-instructAlibaba · Apache 2.0

1401±5

22,789

69

5487

Minimaxminimax-m2.5MiniMax · Modified MIT

1401±9

4,743

70

6379

Anthropicclaude-sonnet-4-20250514-thinking-32kAnthropic · Proprietary

1400±4

36,157

71

6084

longcat-flash-chatMeituan · MIT

1400±6

11,523

72

6387

qwen3-235b-a22b-thinking-2507Alibaba · Apache 2.0

1399±6

9,237

73

6386

deepseek-r1DeepSeek · MIT

1398±5

18,537

74

6392

qwen3-vl-235b-a22b-thinkingAlibaba · Apache 2.0

1395±7

7,958

75

6394

amazon-nova-experimental-chat-12-10Amazon · Proprietary

1394±10

3,706

76

6689

deepseek-v3-0324DeepSeek · MIT

1394±4

46,636

77

6295

Tencenthunyuan-vision-1.5-thinkingTencent · Proprietary

1393±12

2,222

78

6893

mai-1-previewMicrosoft AI · Proprietary

1392±5

18,092

79

7093

o4-mini-2025-04-16OpenAI · Proprietary

1391±4

46,607

80

7094

gpt-5-mini-highOpenAI · Proprietary

1390±5

27,066

81

7094

mimo-v2-flash (non-thinking)Xiaomi · MIT

1390±5

18,406

82

7094

Anthropicclaude-sonnet-4-20250514Anthropic · Proprietary

1390±4

41,586

83

6895

Stepfunstep-3.5-flashStepFun · Apache 2.0

1390±7

7,404

84

7294

Anthropicclaude-3-7-sonnet-20250219-thinking-32kAnthropic · Proprietary

1388±4

39,863

85

7194

o1-previewOpenAI · Proprietary

1388±5

31,120

86

7097

Tencenthunyuan-t1-20250711Tencent · Proprietary

1387±9

4,793

87

7195

mimo-v2-flash (thinking)Xiaomi · MIT

1387±6

10,889

88

7495

qwen3-coder-480b-a35b-instructAlibaba · Apache 2.0

1386±5

26,589

89

7495

Minimaxminimax-m2.1-previewMiniMax · MIT

1386±5

17,143

90

7595

mistral-medium-2505Mistral · Proprietary

1385±5

34,524

91

7697

qwen3-30b-a3b-instruct-2507Alibaba · Apache 2.0

1383±5

24,055

92

7598

Tencenthunyuan-turbos-20250416Tencent · Proprietary

1383±6

11,049

93

7898

gpt-4.1-mini-2025-04-14OpenAI · Proprietary

1382±4

40,485

94

8498

gemini-2.5-flash-lite-preview-09-2025-no-thinkingGoogle · Proprietary

1379±4

47,015

95

75108

glm-4.6vZ.ai · MIT

1378±11

2,790

96

90104

qwen3-235b-a22bAlibaba · Apache 2.0

1375±5

27,142

97

90104

gemini-2.5-flash-lite-preview-06-17-thinkingGoogle · Proprietary

1375±4

33,874

98

92104

qwen2.5-maxAlibaba · Proprietary

1374±4

33,283

99

95105

Anthropicclaude-3-5-sonnet-20241022Anthropic · Proprietary

1373±3

89,445

100

95108

Anthropicclaude-3-7-sonnet-20250219Anthropic · Proprietary

1372±4

44,417

101

95108

glm-4.5-airZ.ai · MIT

1371±4

31,293

102

95111

qwen3-next-80b-a3b-thinkingAlibaba · Apache 2.0

1369±6

13,826

103

95111

Minimaxminimax-m1MiniMax · Apache 2.0

1367±4

36,789

104

95115

glm-4.7-flashZ.ai · MIT

1366±7

9,315

105

99113

gemma-3-27b-itGoogle · Gemma

1365±4

48,652

106

98115

amazon-nova-experimental-chat-11-10Amazon · Proprietary

1365±5

18,080

107

99116

o3-mini-highOpenAI · Proprietary

1364±5

18,584

108

99119

grok-3-mini-highxAI · Proprietary

1363±5

17,505

109

102119

gemini-2.0-flash-001Google · Proprietary

1361±4

44,811

110

102127

deepseek-v3DeepSeek · DeepSeek

1358±5

21,788

111

104128

grok-3-mini-betaxAI · Proprietary

1357±5

23,745

112

102134

intellect-3Prime Intellect · MIT

1356±8

5,307

113

105131

mistral-small-2506Mistral · Apache 2.0

1356±5

18,327

114

108133

gpt-oss-120bOpenAI · Apache 2.0

1354±4

30,895

115

108134

gemini-2.0-flash-lite-preview-02-05Google · Proprietary

1353±4

24,951

116

104135

glm-4.5vZ.ai · MIT

1353±8

4,968

117

110133

Coherecommand-a-03-2025Cohere · CC-BY-NC-4.0

1353±3

57,360

118

110134

gemini-1.5-pro-002Google · Proprietary

1351±3

55,607

119

110136

amazon-nova-experimental-chat-10-20Amazon · Proprietary

1350±6

11,367

120

105146

Tencenthunyuan-turbos-20250226Tencent · Proprietary

1349±12

2,226

121

112135

o3-miniOpenAI · Proprietary

1348±3

58,606

122

108146

amazon-nova-experimental-chat-10-09Amazon · Proprietary

1348±11

2,886

123

107148

Nvidiallama-3.1-nemotron-ultra-253b-v1Nvidia · Nvidia Open Model

1347±12

2,546

124

110141

Minimaxminimax-m2MiniMax · Apache 2.0

1347±8

6,710

125

110141

ling-flash-2.0Ant Group · MIT

1347±7

7,032

126

110145

qwen3-32bAlibaba · Apache 2.0

1347±9

3,932

127

110143

Stepfunstep-3StepFun · Apache 2.0

1347±7

6,585

128

110145

qwen-plus-0125Alibaba · Proprietary

1346±8

5,823

129

115138

gpt-4o-2024-05-13OpenAI · Proprietary

1346±3

112,863

130

112148

glm-4-plus-0111Zhipu · Proprietary

1343±8

5,760

131

118141

Anthropicclaude-3-5-sonnet-20240620Anthropic · Proprietary

1343±3

82,417

132

112150

gemma-3-12b-itGoogle · Gemma

1342±10

3,829

133

113152

Nvidianvidia-llama-3.3-nemotron-super-49b-v1.5Nvidia · Nvidia Open

1341±10

3,429

134

111153

Tencenthunyuan-turbo-0110Tencent · Proprietary

1340±12

2,295

135

120152

gpt-5-nano-highOpenAI · Proprietary

1338±7

8,383

136

121152

nova-2-liteAmazon · Proprietary

1337±6

12,135

137

122149

o1-miniOpenAI · Proprietary

1337±4

51,986

138

122152

qwq-32bAlibaba · Apache 2.0

1336±4

26,094

139

126151

Metallama-3.1-405b-instruct-bf16Meta · Llama 3.1 Community

1335±4

41,392

140

126152

grok-2-2024-08-13xAI · Proprietary

1335±4

63,495

141

124152

gpt-4o-2024-08-06OpenAI · Proprietary

1335±4

45,498

142

122152

gemini-advanced-0514Google · Proprietary

1335±5

50,142

143

121159

Stepfunstep-2-16k-exp-202412StepFun · Proprietary

1334±9

4,829

144

128152

Metallama-3.1-405b-instruct-fp8Meta · Llama 3.1 Community

1334±4

59,655

145

130160

olmo-3.1-32b-instructAllen AI · Apache 2.0

1331±6

12,287

146

132164

01.AIyi-lightning01 AI · Proprietary

1329±5

27,340

147

113183

molmo-2-8bAI2 · Apache 2.0

1328±21

821

148

133165

qwen3-30b-a3bAlibaba · Apache 2.0

1328±5

27,408

149

134165

Metallama-4-maverick-17b-128e-instructMeta · Llama 4

1328±4

41,096

150

125174

Nvidiallama-3.3-nemotron-49b-super-v1Nvidia · Nvidia

1327±12

2,230

151

130174

Tencenthunyuan-large-2025-02-10Tencent · Proprietary

1326±10

3,738

152

144170

gpt-4-turbo-2024-04-09OpenAI · Proprietary

1324±4

98,130

153

144170

Anthropicclaude-3-5-haiku-20241022Anthropic · Proprietary

1324±3

71,152

154

135174

deepseek-v2.5-1210DeepSeek · DeepSeek

1323±8

6,793

155

144170

gemini-1.5-pro-001Google · Proprietary

1323±4

79,132

156

144171

Metallama-4-scout-17b-16e-instructMeta · Llama

1323±5

31,203

157

145170

Anthropicclaude-3-opus-20240229Anthropic · Proprietary

1322±3

194,904

158

143175

gpt-4.1-nano-2025-04-14OpenAI · Proprietary

1322±8

6,107

159

144174

Stepfunstep-1o-turbo-202506StepFun · Proprietary

1322±7

9,683

160

144176

ring-flash-2.0Ant Group · MIT

1320±7

7,190

161

149174

Metallama-3.3-70b-instructMeta · Llama-3.3

1320±3

55,552

162

146174

glm-4-plusZhipu AI · Proprietary

1319±5

26,134

163

146176

gemma-3n-e4b-itGoogle · Gemma

1319±5

23,320

164

146177

qwen-max-0919Alibaba · Qwen

1318±6

16,479

165

149174

gpt-4o-mini-2024-07-18OpenAI · Proprietary

1318±3

68,800

166

147181

gpt-oss-20bOpenAI · Apache 2.0

1317±6

10,791

167

149180

Nvidianvidia-nemotron-3-nano-30b-a3b-bf16Nvidia · NVIDIA Open Model

1317±6

15,444

168

149182

qwen2.5-plus-1127Alibaba · Proprietary

1315±6

10,179

169

153181

athene-v2-chatNexusFlow · NexusFlow

1314±5

24,746

170

153181

mistral-large-2407Mistral · Mistral Research

1314±4

45,460

171

154182

gpt-4-0125-previewOpenAI · Proprietary

1313±4

93,439

172

154182

gpt-4-1106-previewOpenAI · Proprietary

1313±4

100,107

173

149186

Tencenthunyuan-standard-2025-02-10Tencent · Proprietary

1312±10

3,905

174

161185

gemini-1.5-flash-002Google · Proprietary

1310±4

34,909

175

146191

mercuryInception AI · Proprietary

1310±14

1,896

176

165186

grok-2-mini-2024-08-13xAI · Proprietary

1308±4

52,574

177

165186

deepseek-v2.5DeepSeek · DeepSeek

1307±5

24,574

178

165186

athene-70b-0725NexusFlow · CC-BY-NC-4.0

1306±6

19,622

179

162186

olmo-3-32b-thinkAllen AI · Apache 2.0

1306±8

5,878

180

169186

mistral-large-2411Mistral · MRL

1305±4

28,081

181

166186

magistral-medium-2506Mistral · Proprietary

1305±6

12,058

182

171186

mistral-small-3.1-24b-instruct-2503Mistral · Apache 2.0

1305±4

34,082

183

164193

gemma-3-4b-itGoogle · Gemma

1303±9

4,177

184

173186

qwen2.5-72b-instructAlibaba · Qwen

1303±4

39,409

185

173196

Nvidiallama-3.1-nemotron-70b-instructNvidia · Llama 3.1

1299±8

7,136

186

174197

Tencenthunyuan-large-visionTencent · Proprietary

1296±9

5,597

187

183197

Metallama-3.1-70b-instructMeta · Llama 3.1 Community

1294±4

55,234

188

184198

amazon-nova-pro-v1.0Amazon · Proprietary

1290±5

24,753

189

183201

jamba-1.5-largeAI21 Labs · Jamba Open

1289±7

8,659

190

183202

ibm-granite-h-smallIBM · Apache 2.0

1288±8

5,657

191

185199

gemma-2-27b-itGoogle · Gemma license

1288±3

75,764

192

184201

reka-core-20240904Reka AI · Proprietary

1288±7

7,309

193

185201

gpt-4-0314OpenAI · Proprietary

1287±5

54,167

194

183207

llama-3.1-tulu-3-70bAi2 · Llama 3.1

1287±10

2,846

195

183207

Nvidiallama-3.1-nemotron-51b-instructNvidia · Llama 3.1

1286±10

3,749

196

186201

gemini-1.5-flash-001Google · Proprietary

1286±4

62,823

197

185207

olmo-3.1-32b-thinkAllen AI · Apache 2.0

1285±7

8,455

198

189207

Anthropicclaude-3-sonnet-20240229Anthropic · Proprietary

1281±4

109,289

199

188207

gemma-2-9b-it-simpoPrinceton · MIT

1279±7

10,069

200

190207

Nvidianemotron-4-340b-instructNvidia · NVIDIA Open Model

1278±5

19,661

201

190209

Coherecommand-r-plus-08-2024Cohere · CC-BY-NC-4.0

1277±7

9,869

202

195207

Metallama-3-70b-instructMeta · Llama 3 Community

1276±4

156,880

203

195208

gpt-4-0613OpenAI · Proprietary

1275±4

88,721

204

195210

mistral-small-24b-instruct-2501Mistral · Apache 2.0

1274±6

14,677

205

194211

glm-4-0520Zhipu AI · Proprietary

1274±7

9,788

206

195213

reka-flash-20240904Reka AI · Proprietary

1272±7

7,537

207

195216

qwen2.5-coder-32b-instructAlibaba · Apache 2.0

1271±8

5,430

208

202216

Coherec4ai-aya-expanse-32bCohere · CC-BY-NC-4.0

1267±5

27,123

209

204216

gemma-2-9b-itGoogle · Gemma license

1266±4

54,615

210

203217

deepseek-coder-v2DeepSeek · DeepSeek License

1264±6

15,147

211

206217

Coherecommand-r-plusCohere · CC-BY-NC-4.0

1262±4

77,556

212

205217

qwen2-72b-instructAlibaba · Qianwen LICENSE

1262±5

37,325

213

207217

Anthropicclaude-3-haiku-20240307Anthropic · Proprietary

1261±4

117,705

214

206218

amazon-nova-lite-v1.0Amazon · Proprietary

1261±5

19,376

215

207218

gemini-1.5-flash-8b-001Google · Proprietary

1259±4

35,556

216

210218

Azurephi-4Microsoft · MIT

1256±5

24,126

217

207224

olmo-2-0325-32b-instructAllen AI · Apache-2.0

1252±11

3,335

218

214223

Coherecommand-r-08-2024Cohere · CC-BY-NC-4.0

1250±7

10,141

219

217227

mistral-large-2402Mistral · Proprietary

1243±5

62,437

220

217227

amazon-nova-micro-v1.0Amazon · Proprietary

1241±5

19,355

221

217231

jamba-1.5-miniAI21 Labs · Jamba Open

1239±7

8,854

222

217235

ministral-8b-2410Mistral · MRL

1237±9

4,780

223

218235

gemini-pro-dev-apiGoogle · Proprietary

1235±7

18,352

224

219235

qwen1.5-110b-chatAlibaba · Qianwen LICENSE

1234±6

26,191

225

219236

reka-flash-21b-20240226-onlineReka AI · Proprietary

1233±7

15,451

226

217237

Tencenthunyuan-standard-256kTencent · Proprietary

1233±12

2,729

227

219235

qwen1.5-72b-chatAlibaba · Qianwen LICENSE

1233±5

39,296

228

221236

mixtral-8x22b-instruct-v0.1Mistral · Apache 2.0

1230±5

51,417

229

221237

Coherecommand-rCohere · CC-BY-NC-4.0

1227±5

54,038

230

221237

reka-flash-21b-20240226Reka AI · Proprietary

1227±6

24,806

231

222238

gpt-3.5-turbo-0125OpenAI · Proprietary

1224±5

66,191

232

226238

Metallama-3-8b-instructMeta · Llama 3 Community

1223±4

104,636

233

222239

mistral-mediumMistral · Proprietary

1223±6

34,552

234

222239

Coherec4ai-aya-expanse-8bCohere · CC-BY-NC-4.0

1223±7

9,827

235

221242

gemini-proGoogle · Proprietary

1222±12

6,390

236

222242

llama-3.1-tulu-3-8bAi2 · Llama 3.1

1221±11

2,895

237

233242

01.AIyi-1.5-34b-chat01 AI · Apache-2.0

1214±5

24,142

238

228244

HuggingFacezephyr-orpo-141b-A35b-v0.1HuggingFace · Apache 2.0

1213±11

4,653

239

235242

Metallama-3.1-8b-instructMeta · Llama 3.1 Community

1212±4

49,605

240

231248

granite-3.1-8b-instructIBM · Apache 2.0

1209±11

3,092

241

235248

qwen1.5-32b-chatAlibaba · Qianwen LICENSE

1204±6

21,744

242

235250

gpt-3.5-turbo-1106OpenAI · Proprietary

1203±9

16,616

243

239249

gemma-2-2b-itGoogle · Gemma license

1199±4

46,618

244

239250

Azurephi-3-medium-4k-instructMicrosoft · MIT

1198±5

25,055

245

240250

mixtral-8x7b-instruct-v0.1Mistral · Apache 2.0

1197±4

73,505

246

240255

dbrx-instruct-previewDatabricks · DBRX LICENSE

1195±6

32,196

247

240259

InternLMinternlm2_5-20b-chatInternLM · Other

1191±7

9,902

248

240259

qwen1.5-14b-chatAlibaba · Qianwen LICENSE

1191±7

17,841

249

243265

Azurewizardlm-70bMicrosoft · Llama 2 Community

1185±9

8,214

250

242266

deepseek-llm-67b-chatDeepSeek · DeepSeek License

1185±12

4,933

251

246262

01.AIyi-34b-chat01 AI · Yi License

1184±7

15,483

252

246265

OpenChatopenchat-3.5-0106OpenChat · Apache-2.0

1183±8

12,636

253

246266

OpenChatopenchat-3.5OpenChat · Apache-2.0

1182±10

7,967

254

246266

granite-3.0-8b-instructIBM · Apache 2.0

1182±9

6,643

255

247266

gemma-1.1-7b-itGoogle · Gemma license

1180±6

23,893

256

247266

Snowflakesnowflake-arctic-instructSnowflake · Apache 2.0

1180±6

32,836

257

246267

granite-3.1-2b-instructIBM · Apache 2.0

1179±11

3,191

258

247267

tulu-2-dpo-70bAllenAI/UW · AI2 ImpACT Low-risk

1178±10

6,534

259

247270

openhermes-2.5-mistral-7bNousResearch · Apache-2.0

1175±10

5,006

260

249269

vicuna-33bLMSYS · Non-commercial

1173±6

22,479

261

249271

starling-lm-7b-betaNexusflow · Apache-2.0

1172±7

16,057

262

249270

Azurephi-3-small-8k-instructMicrosoft · MIT

1171±6

17,763

263

250270

Metallama-2-70b-chatMeta · Llama 2 Community

1171±6

38,491

264

250273

starling-lm-7b-alphaUC Berkeley · CC-BY-NC-4.0

1168±8

10,224

265

252273

Metallama-3.2-3b-instructMeta · Llama 3.2

1167±8

7,936

266

250276

nous-hermes-2-mixtral-8x7b-dpoNousResearch · Apache-2.0

1165±12

3,776

267

257282

qwq-32b-previewAlibaba · Apache 2.0

1157±12

3,233

268

263280

granite-3.0-2b-instructIBM · Apache 2.0

1156±8

6,837

269

259283

Nvidiallama2-70b-steerlm-chatNvidia · Llama 2 Community

1156±13

3,584

270

260286

solar-10.7b-instruct-v1.0Upstage AI · CC-BY-NC-4.0

1153±13

4,155

271

259288

dolphin-2.2.1-mistral-7bCognitive Computations · Apache-2.0

1152±15

1,679

272

264286

mpt-30b-chatMosaicML · CC-BY-NC-SA-4.0

1150±12

2,571

273

266283

mistral-7b-instruct-v0.2Mistral · Apache-2.0

1150±7

19,402

274

266285

Azurewizardlm-13bMicrosoft · Llama 2 Community

1149±9

7,046

275

264290

falcon-180b-chatTII · Falcon-180B TII License

1147±17

1,295

276

266289

qwen1.5-7b-chatAlibaba · Qianwen LICENSE

1144±10

4,735

277

267288

Azurephi-3-mini-4k-instruct-june-2024Microsoft · MIT

1143±6

12,296

278

267289

Metallama-2-13b-chatMeta · Llama 2 Community

1142±7

19,171

279

267289

vicuna-13bLMSYS · Llama 2 Community

1141±7

19,366

280

267291

qwen-14b-chatAlibaba · Qianwen LICENSE

1139±11

4,964

281

268291

palm-2Google · Proprietary

1137±9

8,554

282

268291

Metacodellama-34b-instructMeta · Llama 2 Community

1137±9

7,363

283

269291

gemma-7b-itGoogle · Gemma license

1136±10

8,925

284

271292

HuggingFacezephyr-7b-betaHuggingFace · MIT

1131±9

11,116

285

274292

Azurephi-3-mini-128k-instructMicrosoft · MIT

1129±7

20,691

286

276292

Azurephi-3-mini-4k-instructMicrosoft · MIT

1129±6

20,115

287

272296

guanaco-33bUW · Non-commercial

1128±12

2,921

288

271296

HuggingFacezephyr-7b-alphaHuggingFace · MIT

1127±16

1,785

289

279296

stripedhyena-nous-7bTogether AI · Apache 2.0

1121±11

5,184

290

274297

Metacodellama-70b-instructMeta · Llama 2 Community

1119±18

1,143

291

284296

vicuna-7bLMSYS · Llama 2 Community

1115±9

6,923

292

280297

HuggingFacesmollm2-1.7b-instructHuggingFace · Apache 2.0

1114±14

2,201

293

287296

gemma-1.1-2b-itGoogle · Gemma license

1114±8

10,853

294

287296

Metallama-3.2-1b-instructMeta · Llama 3.2

1111±8

8,045

295

287297

mistral-7b-instructMistral · Apache 2.0

1110±9

8,977

296

287297

Metallama-2-7b-chatMeta · Llama 2 Community

1108±7

14,148

297

293301

gemma-2b-itGoogle · Gemma license

1092±12

4,779

298

297300

qwen1.5-4b-chatAlibaba · Qianwen LICENSE

1090±9

7,598

299

297304

olmo-7b-instructAllen AI · Apache-2.0

1075±11

6,329

300

298304

koala-13bUC Berkeley · Non-commercial

1071±10

6,964

301

299304

alpaca-13bStanford · Non-commercial

1068±12

5,745

302

297305

gpt4all-13b-snoozyNomic AI · Non-commercial

1066±15

1,743

303

299305

mpt-7b-chatMosaicML · CC-BY-NC-SA-4.0

1062±12

3,925

304

299305

chatglm3-6bTsinghua · Apache-2.0

1056±12

4,658

305

302307

RWKVRWKV-4-Raven-14BRWKV · Apache 2.0

1041±11

4,845

306

305307

chatglm2-6bTsinghua · Apache-2.0

1024±14

2,657

307

305307

oasst-pythia-12bOpenAssistant · Apache 2.0

1022±11

6,311

308

308311

chatglm-6bTsinghua · Non-commercial

996±13

4,914

309

308311

fastchat-t5-3bLMSYS · Apache 2.0

991±12

4,203

310

308311

dolly-v2-12bDatabricks · MIT

980±14

3,412

311

308312

Metallama-13bMeta · Non-commercial

972±16

2,391

312

311312

Stabilitystablelm-tuned-alpha-7bStability AI · CC-BY-NC-SA-4.0

953±13

3,287

说明

  • 排名 (UB):基于 Bradley-Terry 模型计算的排名。此排名反映了模型在竞技场中的综合表现,并提供了其 Elo 分数的 上界 估计,帮助理解模型的潜在竞争力。

  • 模型:大型语言模型 (LLM) 的名称。部分模型名称可能已嵌入相关链接。

  • 分数:模型在竞技场中通过用户投票获得的 Elo 评分。Elo 评分是一种相对排名系统,分数越高表示模型表现越好。

  • 95% 置信区间 (±):模型 Elo 评分的95%置信区间(例如:±6)。这个区间越小,表示模型的评分越稳定和可靠。

  • 票数:该模型在竞技场中收到的总投票数量。投票数越多,通常意味着其评分的统计可靠性越高。

  • 组织/公司:提供该模型的组织或公司。

  • 许可证:模型的许可协议类型,例如专有 (Proprietary)、Apache 2.0、MIT 等。

数据来源与更新频率

本排行榜数据由自动化脚本直接从 1 2arrow-up-right 官方网站获取。此排行榜由 GitHub Actions 每天自动更新。

免责声明

本报告仅供参考。排行榜数据是动态变化的,并基于特定时间段内用户在 Chatbot Arena 上的偏好投票。数据的完整性和准确性取决于上游数据源。不同模型可能采用不同的许可协议,使用时请务必参考模型提供商的官方说明。

最后更新于

这有帮助吗?