No model matches this selection.
- 176models
- 54makers
- 2026-10-01as of (Artificial Analysis)
Understanding the measurements
能力指 Artificial Analysis(AA)的 Intelligence Index,这是一个独立的评测网站:数值越高越好。每项任务成本是 AA 通过 API 对其测试集每项任务测得的美元金额,与 AA 公布的数字一致。速度以每秒 token 数计。* = AA 估算值,并非完整测量。– = 未测量。
全部当前模型
按能力排序。大字为最佳档位,下方为每个档位的能力、每项任务成本和速度。
AnthropicClaude Opus 5.5
58能力(最佳档位)42low$0.5572 t/s51medium$1.3472 t/s54high$1.8274 t/s56xhigh$3.4679 t/s58max$5.9892 t/sAnthropicClaude Sonnet 5.5
56能力(最佳档位)low–88 t/s41medium$0.5992 t/s47high$1.0894 t/s52xhigh$2.74103 t/s56max$7.62139 t/sAnthropicClaude Fable 5.1
53能力(最佳档位)47low$2.3747 t/s49medium$2.9850 t/s51high$3.9151 t/s53xhigh$5.9861 t/s53max$7.6369 t/sOpenAIGPT-6 Astra
53能力(最佳档位)46low$0.8243 t/s50medium$1.5445 t/s51high$1.7345 t/s52xhigh$2.3148 t/s53max$3.2651 t/sGoogleGemini 4 Argon
53能力(最佳档位)53high$1.99–OpenAIGPT-6.1 Sol
52能力(最佳档位)42low$0.1360 t/s48medium$0.2159 t/s50high$0.3263 t/s51xhigh$0.3963 t/s52max$0.7265 t/sMetaMuse Spark 1.3
48能力(最佳档位)45xhigh$1.37142 t/s48max$1.60181 t/sxAIGrok 4.7
46能力(最佳档位)46high$2.7376 t/s46xhigh$3.7473 t/sXiaomiMiMo-V2.6-Pro
46能力(最佳档位)46–$0.1342 t/sZ.aiGLM-5.3
45能力(最佳档位)34low$0.8569 t/s45max$2.0168 t/sAlibabaQwen3.8 Max
45能力(最佳档位)45–$5.4139 t/s- Moonshot AIKimi K344能力(最佳档位)30low$1.1536 t/s44max$2.0034 t/s
StepFunStep 5 Preview
44能力(最佳档位)44–$0.7286 t/sZ.aiGLM-5.3-Flash
42能力(最佳档位)42max$0.2545 t/sGoogleGemini 3.8 Flash
41能力(最佳档位)33low––40medium$0.93–41high$1.24236 t/sAlibabaQwen3.8 2.4T A95B
40能力(最佳档位)40–$2.1640 t/sAlibabaQwen3.8-Flash-Next
40能力(最佳档位)40–$0.3758 t/sDeepSeekDeepSeek V4.1 Flash
39能力(最佳档位)25aus$0.15202 t/s39max$0.27209 t/sXiaomiMiMo-V2.6-Flash
38能力(最佳档位)38–$0.0654 t/sOpenAIGPT-6 Luna
37能力(最佳档位)18aus$0.01132 t/s21low$0.0045134 t/s29medium$0.02–32high$0.03123 t/s34xhigh$0.04131 t/s37max$0.07127 t/sDeepSeekDeepSeek V4 Pro 0813
36能力(最佳档位)20aus$0.48180 t/s36max$0.6781 t/sDeepSeekDeepSeek V4 Flash Vision
35能力(最佳档位)35max$0.31215 t/s- China MobileJT-4.1 Flash 236B A21B34能力(最佳档位)34–––27*aus––
Motif TechnologiesMotif 3
34能力(最佳档位)34*–––AlibabaQwen3.8 27B
34能力(最佳档位)20aus$2.4950 t/s26low$1.0550 t/s28medium$1.1349 t/s34xhigh$1.0144 t/sInstitute of Foundation ModelsK2 Horizon 375B A23B
31能力(最佳档位)31–––GoogleGemini 3.1 Pro Preview
30能力(最佳档位)30–$0.67113 t/sMiniMaxMiniMax-M3
29能力(最佳档位)29–$0.5199 t/sNex AGINex-N2-Pro
28能力(最佳档位)28*–––UpstageSolar Pro 4
28能力(最佳档位)28*medium–80 t/sMultiverse ComputingQuasar 438B
27能力(最佳档位)27max$2.02143 t/sApodexApodex 1.1
26能力(最佳档位)26–$0.46–Thinking MachinesInkling Small
26能力(最佳档位)26high$0.09188 t/s- Moonshot AIKimi K2.7 Code26能力(最佳档位)26–$0.5483 t/s
TencentHy3
25能力(最佳档位)25–$0.0788 t/sThinking MachinesInkling
25能力(最佳档位)25xhigh–160 t/sInstitute of Foundation ModelsK2 Horizon MoVA 36B A4B
25能力(最佳档位)25–––- Ant GroupLing-3.0-flash-VL25能力(最佳档位)25––142 t/s
UpstageSolar Open2 250B
25能力(最佳档位)25*–––UpstageSolar Mini 4
24能力(最佳档位)24–$0.36204 t/sSK TelecomA.X-K2
23能力(最佳档位)23*–––- Ant GroupLing-3.0-flash-Fin23能力(最佳档位)23––303 t/s
NVIDIANemotron 3 Ultra
23能力(最佳档位)23high$0.60180 t/sAI9StarsG9v3-39A5B
22能力(最佳档位)22*–––GoogleGemini 3.5 Flash-Lite
22能力(最佳档位)22–$0.12339 t/sKwaiKATKAT-Coder-Pro V2
22能力(最佳档位)22*–––Institute of Foundation ModelsK2 Horizon 7B
21能力(最佳档位)21–––AlibabaQwen3.5 397B A17B
21能力(最佳档位)18–$0.4789 t/s21*aus–86 t/s- LG AI ResearchK-EXAONE 2.020能力(最佳档位)20*–––
- Ant GroupLing 3.0 Flash20能力(最佳档位)20––325 t/s
AlibabaQwen3.5 Omni Plus
20能力(最佳档位)20*––80 t/s- China MobileJT-35B-Flash19能力(最佳档位)19*–––
- MeituanLongCat 2.019能力(最佳档位)19–$0.06–
AlibabaQwen3.5 122B A10B
18能力(最佳档位)16–$0.32125 t/s18*aus–138 t/sAlibabaQwen3.6 35B A3B
18能力(最佳档位)18–$0.48125 t/s15*aus–139 t/sAnthropicClaude 4.5 Haiku
17能力(最佳档位)17–$0.2891 t/s15*aus–83 t/sByteDance SeedDoubao Seed Code
17能力(最佳档位)17*–––GoogleGemma 4 26B A4B
17能力(最佳档位)17*–––13*aus–112 t/sMetaMuse Glimmer
17能力(最佳档位)17high$0.06148 t/s- Ant GroupRing-2.6-1T17能力(最佳档位)17–$0.29118 t/s
Institute of Foundation ModelsK2 Horizon 3.7B
16能力(最佳档位)16–––GoogleGemma 4 31B
15能力(最佳档位)15–$036 t/s14*aus–40 t/sIBMGranite 4.2 30B
15能力(最佳档位)15*––77 t/sBaiduERNIE 5.0 Thinking Preview
14能力(最佳档位)14*–––GoogleGemma 4 12B
14能力(最佳档位)14*––111 t/s9*aus–117 t/s- Mistral AIMistral Medium 3.514能力(最佳档位)14–$0.50165 t/s
AmazonNova 2.0 Omni
14能力(最佳档位)8*aus––11*low––14*medium––AmazonNova 2.0 Pro Preview
14能力(最佳档位)10*aus–106 t/s13*low–124 t/s14*medium–113 t/s- ServiceNowApriel-v1.6-15B-Thinker13能力(最佳档位)13*–––
CohereCommand A+
13能力(最佳档位)13–$0211 t/s- LG AI ResearchEXAONE 4.5 33B13能力(最佳档位)13*–––aus––
NVIDIANemotron 3 Super
13能力(最佳档位)13medium$1.64165 t/sNVIDIANemotron 3.5 Lightning
13能力(最佳档位)13–$0.09302 t/sAmazonNova 2.0 Lite
13能力(最佳档位)9*aus–169 t/s12*low–192 t/s12*medium–179 t/s13*high–183 t/sAlibabaQwen3.5 4B
13能力(最佳档位)13*––31 t/s11*aus–16 t/sAlibabaQwen3.5 9B
13能力(最佳档位)11–$0.2134 t/s13*aus––Multiverse ComputingHyperNova 60B 2605
12能力(最佳档位)12*high––- China MobileJT-MINI12能力(最佳档位)12*–––
- Mistral AIMagistral Medium 1.212能力(最佳档位)12*–––
InceptionMercury 2.5
12能力(最佳档位)12–$0.12637 t/sOpenBMBMiniCPM5-2B
12能力(最佳档位)12–––NVIDIANemotron Cascade 2 30B A3B
12能力(最佳档位)12*–––AlibabaQwen3.5 Omni Flash
12能力(最佳档位)12*––219 t/sOpenAIgpt-oss-120b
12能力(最佳档位)10*low–171 t/s12high$0.11174 t/sAI9StarsG9v3-3B
11能力(最佳档位)11*–––IBMGranite 4.2 8B
11能力(最佳档位)11–$0.0265 t/s- NaverHyperCLOVA X SEED Think (32B)11能力(最佳档位)11*–––
Prime IntellectINTELLECT-3
11能力(最佳档位)11*–––Institute of Foundation ModelsK2 Think V2
11能力(最佳档位)11*–––- Ant GroupLing 3.0 Tiny11能力(最佳档位)11–$057 t/s
- MeituanLongCat Flash Lite11能力(最佳档位)11*–––
- Korea TelecomMi:dm K 2.5 Pro11能力(最佳档位)11*–––
- Mistral AIMistral Small 411能力(最佳档位)11–$0.02174 t/s9*aus–167 t/s
AlibabaQwen3 Next 80B A3B
11能力(最佳档位)11*––157 t/s10*––165 t/sArcee AITrinity Large Thinking
11能力(最佳档位)11–$0.12336 t/sGoogleDiffusionGemma 26B A4B
10能力(最佳档位)10*–––Institute of Foundation ModelsK2-V2
10能力(最佳档位)7*low––9*medium––10*high––MetaLlama 4 Maverick
10能力(最佳档位)10*––84 t/sNVIDIANemotron 3 Nano Omni 30B A3B
10能力(最佳档位)10*––264 t/sCohereNorth Mini Code
10能力(最佳档位)10–$070 t/sUpstageSolar Open 100B
10能力(最佳档位)10*–––OpenAIgpt-oss-20b
10能力(最佳档位)10*low–198 t/s9high$0.01163 t/sGoogleGemma 4 E4B
9能力(最佳档位)9*––70 t/s7*aus–72 t/sIBMGranite 4.2 3B
9能力(最佳档位)9–$0.01223 t/s- Mistral AIMagistral Small 1.29能力(最佳档位)9*–––
OpenBMBMiniCPM5-1B
9能力(最佳档位)9*–––9*aus––- Mistral AIMistral Large 39能力(最佳档位)9–$0.0379 t/s
Motif TechnologiesMotif-2-12.7B
9能力(最佳档位)9*–––NVIDIANemotron 3 Nano
9能力(最佳档位)9–$0.02197 t/s7*aus–231 t/sAmazonNova Premier
9能力(最佳档位)9*–––AlibabaQwen3 Coder Next
9能力(最佳档位)9–$0.5575 t/s- SarvamSarvam 105B9能力(最佳档位)9*high––
Trillion LabsTri-21B-Think
9能力(最佳档位)9*–––BaiduERNIE 4.5 300B A47B
8能力(最佳档位)8*–––- LG AI ResearchEXAONE 4.0 32B8能力(最佳档位)8*–––6*aus––
TII UAEFalcon-H1R-7B
8能力(最佳档位)8*–––GoogleGemma 4 E2B
8能力(最佳档位)8*–––6*aus––- Nous ResearchHermes 4 70B8能力(最佳档位)8*–––7*aus––
Liquid AILFM2.5-2.6B
8能力(最佳档位)8*–––MetaLlama 3.3 70B
8能力(最佳档位)8*––82 t/sMetaLlama 4 Scout
8能力(最佳档位)8*––129 t/sNVIDIALlama Nemotron Ultra
8能力(最佳档位)8*–––- NanbeigeNanbeige4.1-3B8能力(最佳档位)8*–––
AlibabaQwen3 Omni 30B A3B
8能力(最佳档位)8*––102 t/s6*––96 t/sStepFunStep3 VL 10B
8能力(最佳档位)8*–––CohereCommand A
7能力(最佳档位)7*––62 t/s- Nous ResearchHermes 4 405B7能力(最佳档位)7*––38 t/s7*aus–39 t/s
- Moonshot AIKimi Linear 48B A3B Instruct7能力(最佳档位)7*–––
Liquid AILFM2.5-8B-A1B
7能力(最佳档位)7*–––MetaLlama 3.1 405B
7能力(最佳档位)7*–––NVIDIANVIDIA Nemotron Nano 12B v2 VL
7能力(最佳档位)7*–––6*aus–202 t/sNVIDIANVIDIA Nemotron Nano 9B V2
7能力(最佳档位)7*––104 t/s7*aus–150 t/sNVIDIANemotron 3 Nano 4B
7能力(最佳档位)7*–––Allen Institute for AIOlmo 3.1 32B Think
7能力(最佳档位)7*–––AlibabaQwen3.5 2B
7能力(最佳档位)7*–––6*aus––- Ant GroupRing-flash-2.07能力(最佳档位)7*–––
- SarvamSarvam 30B7能力(最佳档位)7*high––
- CelerisCeleris-16能力(最佳档位)6–$0.051535 t/s
- Nous ResearchDeepHermes 3 - Mistral 24B6能力(最佳档位)6*aus––
IBMGranite 4.0 H Small
6能力(最佳档位)6*––12 t/sAI21 LabsJamba 1.7 Large
6能力(最佳档位)6*–––AI21 LabsJamba Reasoning 3B
6能力(最佳档位)6*–––Liquid AILFM2 24B A2B
6能力(最佳档位)6*–––MetaLlama 3.2 90B (Vision)
6能力(最佳档位)6*–––OpenBMBMiniCPM-V 4.6 1.3B
6能力(最佳档位)6*–––- Mistral AIMinistral 3 14B6能力(最佳档位)6–$0.0274 t/s
Allen Institute for AIMolmo 7B-D
6能力(最佳档位)6*–––AmazonNova Micro
6能力(最佳档位)6*––249 t/sAllen Institute for AIOlmo 3 7B Think
6能力(最佳档位)6*–––Allen Institute for AIOlmo 3.1 32B Instruct
6能力(最佳档位)6*–––MicrosoftPhi-4
6能力(最佳档位)6*––40 t/sMicrosoftPhi-4 Mini
6能力(最佳档位)6*––45 t/sMicrosoftPhi-4 Multimodal
6能力(最佳档位)6*–––AlibabaQwen3.5 0.8B
6能力(最佳档位)6*–––5*aus––PerplexityR1 1776
6能力(最佳档位)6*–––Reka AIReka Flash 3
6能力(最佳档位)6*–––- Swiss AI InitiativeApertus 70B Instruct5能力(最佳档位)5*–––
- Swiss AI InitiativeApertus 8B Instruct5能力(最佳档位)5*–––
- Nous ResearchDeepHermes 3 - Llama-3.1 8B5能力(最佳档位)5*aus––
- LG AI ResearchExaone 4.0 1.2B5能力(最佳档位)5*–––5*aus––
GoogleGemma 3 270M
5能力(最佳档位)5*–––IBMGranite 4.0 350M
5能力(最佳档位)5*–––IBMGranite 4.0 H 1B
5能力(最佳档位)5*–––IBMGranite 4.0 H 350M
5能力(最佳档位)5*–––IBMGranite 4.0 Micro
5能力(最佳档位)5*–––AI21 LabsJamba 1.7 Mini
5能力(最佳档位)5*–––Liquid AILFM2.5-1.2B-Instruct
5能力(最佳档位)5*–––Liquid AILFM2.5-1.2B-Thinking
5能力(最佳档位)5*–––Liquid AILFM2.5-VL-1.6B
5能力(最佳档位)5*–––MetaLlama 3.2 11B (Vision)
5能力(最佳档位)5*––15 t/s- Mistral AIMinistral 3 3B5能力(最佳档位)5–$0.01198 t/s
- Mistral AIMinistral 3 8B5能力(最佳档位)5–$0.0196 t/s
Allen Institute for AIMolmo2-8B
5能力(最佳档位)5*–––Allen Institute for AIOlmo 3 7B
5能力(最佳档位)5*–––CohereTiny Aya Global
5能力(最佳档位)5*–––Institute of Foundation ModelsK2 Horizon 0.9B
3能力(最佳档位)3–––
全部模型数据
同样的模型以表格列出;成本和速度为各已测档位的范围。
| 模型 | 厂商 | 能力 | 档位(已测) | 每项任务成本 | 速度(token/秒) |
|---|---|---|---|---|---|
| Claude Opus 5.5 | Anthropic | 58 | low · medium · high · xhigh · max | $0.55 – $5.98 | 72 – 92 |
| Claude Sonnet 5.5 | Anthropic | 56 | low · medium · high · xhigh · max | $0.59 – $7.62 | 88 – 139 |
| Claude Fable 5.1 | Anthropic | 53 | low · medium · high · xhigh · max | $2.37 – $7.63 | 47 – 69 |
| GPT-6 Astra | OpenAI | 53 | low · medium · high · xhigh · max | $0.82 – $3.26 | 43 – 51 |
| Gemini 4 Argon | 53 | high | $1.99 | – | |
| GPT-6.1 Sol | OpenAI | 52 | low · medium · high · xhigh · max | $0.13 – $0.72 | 59 – 65 |
| Muse Spark 1.3 | Meta | 48 | xhigh · max | $1.37 – $1.60 | 142 – 181 |
| Grok 4.7 | xAI | 46 | high · xhigh | $2.73 – $3.74 | 73 – 76 |
| MiMo-V2.6-Pro | Xiaomi | 46 | – | $0.13 | 42 |
| GLM-5.3 | Z.ai | 45 | low · max | $0.85 – $2.01 | 68 – 69 |
| Qwen3.8 Max | Alibaba | 45 | – | $5.41 | 39 |
| Kimi K3 | Moonshot AI | 44 | low · max | $1.15 – $2.00 | 34 – 36 |
| Step 5 Preview | StepFun | 44 | – | $0.72 | 86 |
| GLM-5.3-Flash | Z.ai | 42 | max | $0.25 | 45 |
| Gemini 3.8 Flash | 41 | low · medium · high | $0.93 – $1.24 | 236 | |
| Qwen3.8 2.4T A95B | Alibaba | 40 | – | $2.16 | 40 |
| Qwen3.8-Flash-Next | Alibaba | 40 | – | $0.37 | 58 |
| DeepSeek V4.1 Flash | DeepSeek | 39 | aus · max | $0.15 – $0.27 | 202 – 209 |
| MiMo-V2.6-Flash | Xiaomi | 38 | – | $0.06 | 54 |
| GPT-6 Luna | OpenAI | 37 | aus · low · medium · high · xhigh · max | $0.0045 – $0.07 | 123 – 134 |
| DeepSeek V4 Pro 0813 | DeepSeek | 36 | aus · max | $0.48 – $0.67 | 81 – 180 |
| DeepSeek V4 Flash Vision | DeepSeek | 35 | max | $0.31 | 215 |
| JT-4.1 Flash 236B A21B | China Mobile | 34 | – · aus | – | – |
| Motif 3 | Motif Technologies | 34 | – | – | – |
| Qwen3.8 27B | Alibaba | 34 | aus · low · medium · xhigh | $1.01 – $2.49 | 44 – 50 |
| K2 Horizon 375B A23B | Institute of Foundation Models | 31 | – | – | – |
| Gemini 3.1 Pro Preview | 30 | – | $0.67 | 113 | |
| MiniMax-M3 | MiniMax | 29 | – | $0.51 | 99 |
| Nex-N2-Pro | Nex AGI | 28 | – | – | – |
| Solar Pro 4 | Upstage | 28 | medium | – | 80 |
| Quasar 438B | Multiverse Computing | 27 | max | $2.02 | 143 |
| Apodex 1.1 | Apodex | 26 | – | $0.46 | – |
| Inkling Small | Thinking Machines | 26 | high | $0.09 | 188 |
| Kimi K2.7 Code | Moonshot AI | 26 | – | $0.54 | 83 |
| Hy3 | Tencent | 25 | – | $0.07 | 88 |
| Inkling | Thinking Machines | 25 | xhigh | – | 160 |
| K2 Horizon MoVA 36B A4B | Institute of Foundation Models | 25 | – | – | – |
| Ling-3.0-flash-VL | Ant Group | 25 | – | – | 142 |
| Solar Open2 250B | Upstage | 25 | – | – | – |
| Solar Mini 4 | Upstage | 24 | – | $0.36 | 204 |
| A.X-K2 | SK Telecom | 23 | – | – | – |
| Ling-3.0-flash-Fin | Ant Group | 23 | – | – | 303 |
| Nemotron 3 Ultra | NVIDIA | 23 | high | $0.60 | 180 |
| G9v3-39A5B | AI9Stars | 22 | – | – | – |
| Gemini 3.5 Flash-Lite | 22 | – | $0.12 | 339 | |
| KAT-Coder-Pro V2 | KwaiKAT | 22 | – | – | – |
| K2 Horizon 7B | Institute of Foundation Models | 21 | – | – | – |
| Qwen3.5 397B A17B | Alibaba | 21 | – · aus | $0.47 | 86 – 89 |
| K-EXAONE 2.0 | LG AI Research | 20 | – | – | – |
| Ling 3.0 Flash | Ant Group | 20 | – | – | 325 |
| Qwen3.5 Omni Plus | Alibaba | 20 | – | – | 80 |
| JT-35B-Flash | China Mobile | 19 | – | – | – |
| LongCat 2.0 | Meituan | 19 | – | $0.06 | – |
| Qwen3.5 122B A10B | Alibaba | 18 | – · aus | $0.32 | 125 – 138 |
| Qwen3.6 35B A3B | Alibaba | 18 | – · aus | $0.48 | 125 – 139 |
| Claude 4.5 Haiku | Anthropic | 17 | – · aus | $0.28 | 83 – 91 |
| Doubao Seed Code | ByteDance Seed | 17 | – | – | – |
| Gemma 4 26B A4B | 17 | – · aus | – | 112 | |
| Muse Glimmer | Meta | 17 | high | $0.06 | 148 |
| Ring-2.6-1T | Ant Group | 17 | – | $0.29 | 118 |
| K2 Horizon 3.7B | Institute of Foundation Models | 16 | – | – | – |
| Gemma 4 31B | 15 | – · aus | $0 | 36 – 40 | |
| Granite 4.2 30B | IBM | 15 | – | – | 77 |
| ERNIE 5.0 Thinking Preview | Baidu | 14 | – | – | – |
| Gemma 4 12B | 14 | – · aus | – | 111 – 117 | |
| Mistral Medium 3.5 | Mistral AI | 14 | – | $0.50 | 165 |
| Nova 2.0 Omni | Amazon | 14 | aus · low · medium | – | – |
| Nova 2.0 Pro Preview | Amazon | 14 | aus · low · medium | – | 106 – 124 |
| Apriel-v1.6-15B-Thinker | ServiceNow | 13 | – | – | – |
| Command A+ | Cohere | 13 | – | $0 | 211 |
| EXAONE 4.5 33B | LG AI Research | 13 | – | – | – |
| Nemotron 3 Super | NVIDIA | 13 | medium | $1.64 | 165 |
| Nemotron 3.5 Lightning | NVIDIA | 13 | – | $0.09 | 302 |
| Nova 2.0 Lite | Amazon | 13 | aus · low · medium · high | – | 169 – 192 |
| Qwen3.5 4B | Alibaba | 13 | – · aus | – | 16 – 31 |
| Qwen3.5 9B | Alibaba | 13 | – · aus | $0.21 | 34 |
| HyperNova 60B 2605 | Multiverse Computing | 12 | high | – | – |
| JT-MINI | China Mobile | 12 | – | – | – |
| Magistral Medium 1.2 | Mistral AI | 12 | – | – | – |
| Mercury 2.5 | Inception | 12 | – | $0.12 | 637 |
| MiniCPM5-2B | OpenBMB | 12 | – | – | – |
| Nemotron Cascade 2 30B A3B | NVIDIA | 12 | – | – | – |
| Qwen3.5 Omni Flash | Alibaba | 12 | – | – | 219 |
| gpt-oss-120b | OpenAI | 12 | low · high | $0.11 | 171 – 174 |
| G9v3-3B | AI9Stars | 11 | – | – | – |
| Granite 4.2 8B | IBM | 11 | – | $0.02 | 65 |
| HyperCLOVA X SEED Think (32B) | Naver | 11 | – | – | – |
| INTELLECT-3 | Prime Intellect | 11 | – | – | – |
| K2 Think V2 | Institute of Foundation Models | 11 | – | – | – |
| Ling 3.0 Tiny | Ant Group | 11 | – | $0 | 57 |
| LongCat Flash Lite | Meituan | 11 | – | – | – |
| Mi:dm K 2.5 Pro | Korea Telecom | 11 | – | – | – |
| Mistral Small 4 | Mistral AI | 11 | – · aus | $0.02 | 167 – 174 |
| Qwen3 Next 80B A3B | Alibaba | 11 | – · – | – | 157 – 165 |
| Trinity Large Thinking | Arcee AI | 11 | – | $0.12 | 336 |
| DiffusionGemma 26B A4B | 10 | – | – | – | |
| K2-V2 | Institute of Foundation Models | 10 | low · medium · high | – | – |
| Llama 4 Maverick | Meta | 10 | – | – | 84 |
| Nemotron 3 Nano Omni 30B A3B | NVIDIA | 10 | – | – | 264 |
| North Mini Code | Cohere | 10 | – | $0 | 70 |
| Solar Open 100B | Upstage | 10 | – | – | – |
| gpt-oss-20b | OpenAI | 10 | low · high | $0.01 | 163 – 198 |
| Gemma 4 E4B | 9 | – · aus | – | 70 – 72 | |
| Granite 4.2 3B | IBM | 9 | – | $0.01 | 223 |
| Magistral Small 1.2 | Mistral AI | 9 | – | – | – |
| MiniCPM5-1B | OpenBMB | 9 | – · aus | – | – |
| Mistral Large 3 | Mistral AI | 9 | – | $0.03 | 79 |
| Motif-2-12.7B | Motif Technologies | 9 | – | – | – |
| Nemotron 3 Nano | NVIDIA | 9 | – · aus | $0.02 | 197 – 231 |
| Nova Premier | Amazon | 9 | – | – | – |
| Qwen3 Coder Next | Alibaba | 9 | – | $0.55 | 75 |
| Sarvam 105B | Sarvam | 9 | high | – | – |
| Tri-21B-Think | Trillion Labs | 9 | – | – | – |
| ERNIE 4.5 300B A47B | Baidu | 8 | – | – | – |
| EXAONE 4.0 32B | LG AI Research | 8 | – · aus | – | – |
| Falcon-H1R-7B | TII UAE | 8 | – | – | – |
| Gemma 4 E2B | 8 | – · aus | – | – | |
| Hermes 4 70B | Nous Research | 8 | – · aus | – | – |
| LFM2.5-2.6B | Liquid AI | 8 | – | – | – |
| Llama 3.3 70B | Meta | 8 | – | – | 82 |
| Llama 4 Scout | Meta | 8 | – | – | 129 |
| Llama Nemotron Ultra | NVIDIA | 8 | – | – | – |
| Nanbeige4.1-3B | Nanbeige | 8 | – | – | – |
| Qwen3 Omni 30B A3B | Alibaba | 8 | – · – | – | 96 – 102 |
| Step3 VL 10B | StepFun | 8 | – | – | – |
| Command A | Cohere | 7 | – | – | 62 |
| Hermes 4 405B | Nous Research | 7 | – · aus | – | 38 – 39 |
| Kimi Linear 48B A3B Instruct | Moonshot AI | 7 | – | – | – |
| LFM2.5-8B-A1B | Liquid AI | 7 | – | – | – |
| Llama 3.1 405B | Meta | 7 | – | – | – |
| NVIDIA Nemotron Nano 12B v2 VL | NVIDIA | 7 | – · aus | – | 202 |
| NVIDIA Nemotron Nano 9B V2 | NVIDIA | 7 | – · aus | – | 104 – 150 |
| Nemotron 3 Nano 4B | NVIDIA | 7 | – | – | – |
| Olmo 3.1 32B Think | Allen Institute for AI | 7 | – | – | – |
| Qwen3.5 2B | Alibaba | 7 | – · aus | – | – |
| Ring-flash-2.0 | Ant Group | 7 | – | – | – |
| Sarvam 30B | Sarvam | 7 | high | – | – |
| Celeris-1 | Celeris | 6 | – | $0.05 | 1,535 |
| DeepHermes 3 - Mistral 24B | Nous Research | 6 | aus | – | – |
| Granite 4.0 H Small | IBM | 6 | – | – | 12 |
| Jamba 1.7 Large | AI21 Labs | 6 | – | – | – |
| Jamba Reasoning 3B | AI21 Labs | 6 | – | – | – |
| LFM2 24B A2B | Liquid AI | 6 | – | – | – |
| Llama 3.2 90B (Vision) | Meta | 6 | – | – | – |
| MiniCPM-V 4.6 1.3B | OpenBMB | 6 | – | – | – |
| Ministral 3 14B | Mistral AI | 6 | – | $0.02 | 74 |
| Molmo 7B-D | Allen Institute for AI | 6 | – | – | – |
| Nova Micro | Amazon | 6 | – | – | 249 |
| Olmo 3 7B Think | Allen Institute for AI | 6 | – | – | – |
| Olmo 3.1 32B Instruct | Allen Institute for AI | 6 | – | – | – |
| Phi-4 | Microsoft | 6 | – | – | 40 |
| Phi-4 Mini | Microsoft | 6 | – | – | 45 |
| Phi-4 Multimodal | Microsoft | 6 | – | – | – |
| Qwen3.5 0.8B | Alibaba | 6 | – · aus | – | – |
| R1 1776 | Perplexity | 6 | – | – | – |
| Reka Flash 3 | Reka AI | 6 | – | – | – |
| Apertus 70B Instruct | Swiss AI Initiative | 5 | – | – | – |
| Apertus 8B Instruct | Swiss AI Initiative | 5 | – | – | – |
| DeepHermes 3 - Llama-3.1 8B | Nous Research | 5 | aus | – | – |
| Exaone 4.0 1.2B | LG AI Research | 5 | – · aus | – | – |
| Gemma 3 270M | 5 | – | – | – | |
| Granite 4.0 350M | IBM | 5 | – | – | – |
| Granite 4.0 H 1B | IBM | 5 | – | – | – |
| Granite 4.0 H 350M | IBM | 5 | – | – | – |
| Granite 4.0 Micro | IBM | 5 | – | – | – |
| Jamba 1.7 Mini | AI21 Labs | 5 | – | – | – |
| LFM2.5-1.2B-Instruct | Liquid AI | 5 | – | – | – |
| LFM2.5-1.2B-Thinking | Liquid AI | 5 | – | – | – |
| LFM2.5-VL-1.6B | Liquid AI | 5 | – | – | – |
| Llama 3.2 11B (Vision) | Meta | 5 | – | – | 15 |
| Ministral 3 3B | Mistral AI | 5 | – | $0.01 | 198 |
| Ministral 3 8B | Mistral AI | 5 | – | $0.01 | 96 |
| Molmo2-8B | Allen Institute for AI | 5 | – | – | – |
| Olmo 3 7B | Allen Institute for AI | 5 | – | – | – |
| Tiny Aya Global | Cohere | 5 | – | – | – |
| K2 Horizon 0.9B | Institute of Foundation Models | 3 | – | – | – |
Language and review status
Main text: ZH · includes English sections. Review notes apply only to the content they name.
Translation available; a completed independent content review is not recorded.
See language coverage and data snapshot →