LLM Leaderboard
Compare quality, price, and speed across today's leading models
Top Klu Index100Kimi K3.1 Pro
Best value$0.26MiniMax-M3.1-Flash
Fastest output1521 tok/sCeleris-1
Fastest first token0.5sCommand A+
Largest context2M tokensGrok 4.20 0309
Most expensive$20 / 1MClaude Fable 5.1
Klu Frontier IndexPrice per 1M tokens, log scale, cheaper to the right
FrontierAnthropicOpenAIGooglexAIMetaMistralOther
Frontier explorerKlu Index against cost per task, a point per reasoning effort
ExploreKlu Index
195 models
- All labs195
- Kimi5
- Anthropic11
- Meta8
- DeepSeek7
- Google14
- StepFun5
- OpenAI15
- Z AI8
- xAI7
- Xiaomi7
- MiniMax4
- Alibaba23
- InclusionAI9
- Mistral3
- Institute of Foundation Models5
- China Mobile3
- NVIDIA8
- Upstage4
- Thinking Machines2
- Cohere3
- OpenBMB3
- IBM6
- Liquid AI6
- Other labs29
| # | Model | Klu Index | $ / 1M tokens | Tokens / s | First token | Context | Frontier |
|---|---|---|---|---|---|---|---|
| 1 | Kimi K3.1 ProKimi | 100 | $4.00 | 46 | 3.0s | 1M | Frontier |
| 2 | Claude Opus 5.5Anthropic | 98 | $8.00 | 98 | 712s | 1M | |
| 3 | Claude Sonnet 5.5Anthropic | 96 | $4.00 | 141 | 462s | 1M | |
| 4 | Muse WonderMeta Model API and Muse Code | 96 | $8.00 | 202 | 31s | 1M | |
| 5 | Muse Wonder FullMuse Code subscription | 96 | Subscription | — | — | 1M | |
| 6 | DeepSeek V4.1 Flash 1017DeepSeek | 93 | $0.53 | 169 | 1.2s | 1M | Frontier |
| 7 | Claude Fable 5.1Anthropic | 92 | $20.00 | 70 | 240s | 1M | |
| 8 | Gemini 4 ArgonGoogle | 91 | $4.00 | 80 | 48s | 1M | |
| 9 | Claude Opus 5Anthropic | 90 | $10.00 | 52 | 73s | 1M | |
| 10 | Muse Spark 1.3Meta | 89 | $2.00 | 159 | 44s | 1M | |
| 11 | Step 5 PreviewStepFun | 89 | $1.43 | 198 | 1.3s | 1M | |
| 12 | GPT-6 AstraOpenAI | 88 | $20.00 | 68 | 353s | 1M | |
| 13 | Claude Fable 5Anthropic | 87 | $20.00 | 68 | 102s | 1M | |
| 14 | GLM-5.3Z AI | 87 | $2.15 | 94 | 2.4s | 1M | |
| 15 | GPT-6.1 SolOpenAI | 87 | $4.00 | 69 | 284s | 1M | |
| 16 | Grok 4.7xAI | 87 | $3.00 | 92 | 76s | 500k | |
| 17 | MiMo-V2.6-ProXiaomi | 87 | $0.54 | 56 | 3.4s | 1M | |
| 18 | MiniMax-M3.1-FlashMiniMax | 87 | $0.26 | 116 | 1.1s | 1M | Frontier |
| 19 | Qwen3.8 Max (0902)Alibaba | 86 | $3.00 | 41 | 2.9s | 984k | |
| 20 | DeepSeek V4.1 FlashDeepSeek | 85 | $0.52 | 177 | 1.0s | 1M | |
| 21 | Ling 3.1 FlashInclusionAI | 85 | $0.45 | 213 | 1.8s | 1M | |
| 22 | Ling 3.1 Flash 1019InclusionAI | 85 | $0.45 | — | — | 1M | |
| 23 | Grok 4.6xAI | 84 | $3.00 | 81 | 34s | 500k | |
| 24 | Kimi K3Kimi | 84 | $6.00 | 54 | 2.9s | 1M | |
| 25 | GPT-5.6 SolOpenAI | 83 | $8.00 | 82 | 108s | 1M | |
| 26 | GPT-6 SolOpenAI | 83 | $4.00 | 123 | 120s | 1.05M | |
| 27 | GLM-5.3-FlashZ AI | 82 | $0.24 | 73 | 2.8s | 1M | Frontier |
| 28 | Muse Spark OpenMeta | 82 | $2.00 | 204 | 0.8s | 1M | |
| 29 | GPT-6.1 TerraOpenAI | 81 | $2.00 | 91 | 132s | 1M | |
| 30 | Muse Spark 1.2Meta | 81 | $2.00 | 292 | 12s | 1M | |
| 31 | Qwen3.8-Flash-NextAlibaba | 81 | $0.23 | 52 | 2.1s | 256k | Frontier |
| 32 | Claude Haiku 5.5Anthropic | 79 | $0.20 | 234 | 353s | 1M | Frontier |
| 33 | Qwen3.8 2.4T A95BAlibaba | 79 | $3.00 | 43 | 3.0s | 262k | |
| 34 | DeepSeek V4 Flash VisionDeepSeek | 78 | $0.66 | 181 | 1.6s | 1M | |
| 35 | Mistral Large 4 PreviewMistral API public preview | 78 | $2.07 | 116 | 1.6s | 512k | |
| 36 | Qwen3.8 MaxAlibaba | 78 | $3.00 | 42 | 2.7s | 1M | |
| 37 | Gemini 3.8 FlashGoogle | 77 | $1.50 | 134 | 28s | 1M | |
| 38 | MiMo-V2.6-FlashXiaomi | 77 | $0.18 | 56 | 3.3s | 1M | Frontier |
| 39 | Claude Opus 4.7Anthropic | 76 | $10.00 | 52 | 25s | 1M | |
| 40 | DeepSeek V4 Pro 0813DeepSeek | 76 | $1.98 | 101 | 1.8s | 1M | |
| 41 | Gemini 3.7 FlashGoogle | 76 | $1.50 | — | — | 1M | |
| 42 | Agnes 2.5 Pro BetaSapiens AI | 75 | $0.15 | 114 | 1.4s | 1M | Frontier |
| 43 | Claude Opus 4.8Anthropic | 75 | $10.00 | 60 | 40s | 1M | |
| 44 | DeepSeek V4 Flash 0731DeepSeek | 75 | $0.18 | 149 | 1.3s | 1M | |
| 45 | GPT-5.6 TerraOpenAI | 75 | $4.50 | 135 | 110s | 1M | |
| 46 | Grok 4.5xAI | 74 | $3.00 | 65 | 12s | 500k | |
| 47 | K2 Horizon 375B A23BInstitute of Foundation Models | 74 | — | 919 | 2.4s | 512k | |
| 48 | GPT-6 LunaOpenAI | 71 | $0.20 | 127 | 94s | 1M | |
| 49 | Beam 1.0Reflection | 70 | $0.60 | 366 | 0.6s | 1M | |
| 50 | GPT-5.6 LunaOpenAI | 70 | $0.45 | 126 | 110s | 1M | |
| 51 | JT-4.1 Flash 236B A21BChina Mobile | 70 | — | — | — | 256k | |
| 52 | Claude Sonnet 5Anthropic | 69 | $4.00 | 84 | 199s | 1M | |
| 53 | GLM-5.2Z AI | 69 | $2.15 | 100 | 3.5s | 1M | |
| 54 | GPT-5.4OpenAI | 69 | $5.63 | 117 | 148s | 1.05M | |
| 55 | GPT-5.5OpenAI | 69 | $11.25 | 89 | 76s | 922k | |
| 56 | Qwen3.8 27BAlibaba | 69 | $1.13 | 42 | 3.9s | 256k | |
| 57 | Gemini 3.6 FlashGoogle | 68 | $1.50 | 197 | 18s | 1M | |
| 58 | Motif 3Motif Technologies | 68 | — | — | — | 256k | |
| 59 | Muse Spark 1.1Meta | 68 | $2.00 | 290 | 13s | 1M | |
| 60 | Motif 3 (Beta)Motif Technologies | 66 | — | — | — | 256k | |
| 61 | Gemini 3.5 FlashGoogle | 65 | $3.38 | 195 | 15s | 1M | |
| 62 | Nemotron 3.5 Super VLNVIDIA | 65 | $0.31 | 449 | 0.8s | 1M | |
| 63 | MiniMax-M3MiniMax | 64 | $0.53 | 105 | 1.3s | 1M | |
| 64 | DeepSeek V4 Pro 0424DeepSeek | 63 | $0.54 | 105 | 1.7s | 1M | |
| 65 | Qwen3.7 MaxAlibaba | 63 | $3.75 | 203 | 2.2s | 1M | |
| 66 | Muse SparkMeta | 61 | — | — | — | 256k | |
| 67 | Quasar 438BMultiverse Computing | 61 | $0.17 | 170 | 1.2s | 1M | |
| 68 | Apodex 1.1Apodex | 60 | $0.98 | — | — | 256k | |
| 69 | Solar Pro 4Upstage | 60 | $0.53 | 89 | 2.0s | 512k | |
| 70 | Claude Opus 4.6Anthropic | 59 | $10.00 | 45 | 27s | 1M | |
| 71 | GLM-5Z AI | 59 | $1.55 | 73 | 1.7s | 200k | |
| 72 | Grok Build 0.1 0616xAI | 59 | $1.25 | 89 | 1.3s | 256k | |
| 73 | Inkling SmallThinking Machines | 59 | $0.53 | 224 | 1.6s | 1M | |
| 74 | MiMo-V2-ProXiaomi | 59 | — | — | — | 1M | |
| 75 | GLM-5-TurboZ AI | 58 | — | — | — | 200k | |
| 76 | K2 Horizon MoVA 36B A4BInstitute of Foundation Models | 58 | — | — | — | 512k | |
| 77 | Nex-N2-ProNex AGI | 58 | — | — | — | 262k | |
| 78 | Qwen3.6 Max PreviewAlibaba | 58 | — | — | — | 256k | |
| 79 | GLM-5.1Z AI | 57 | $1.98 | 75 | 1.7s | 200k | |
| 80 | GPT-5.3 CodexOpenAI | 57 | $4.81 | 105 | 56s | 400k | |
| 81 | InklingThinking Machines | 57 | $1.76 | 195 | 1.8s | 1M | |
| 82 | Claude Sonnet 4.6Anthropic | 55 | $6.00 | 58 | 155s | 1M | |
| 83 | GPT-5.5 Instant (June 2026)OpenAI | 55 | $11.25 | 158 | 0.9s | 400k | |
| 84 | Grok 4.20 0309xAI | 55 | $3.00 | — | — | 2M | |
| 85 | Kimi K2.6Kimi | 55 | $1.71 | 96 | 2.6s | 256k | |
| 86 | Kimi K2.7 CodeKimi | 55 | $1.71 | 83 | 2.6s | 256k | |
| 87 | MiMo-V2-Omni-0327Xiaomi | 55 | — | — | — | 256k | |
| 88 | MiMo-V2.5-ProXiaomi | 55 | $0.56 | 52 | 1.1s | 1M | |
| 89 | Nemotron 3 Ultra 550B A55BNVIDIA | 55 | $1.13 | 198 | 1.1s | 256k | |
| 90 | Qwen3.6 PlusAlibaba | 55 | $1.13 | 55 | 2.2s | 1M | |
| 91 | Hy3Tencent | 54 | $0.24 | 81 | 2.7s | 256k | |
| 92 | Ling-3.0-flash-VLInclusionAI | 54 | — | — | — | 256k | |
| 93 | Solar Mini 4Upstage | 54 | $0.18 | 118 | 1.8s | 1M | |
| 94 | DeepSeek V4 Flash 0420DeepSeek | 53 | $0.14 | 61 | 1.3s | 1M | Frontier |
| 95 | Solar Open2 250BUpstage | 53 | — | — | — | 1M | |
| 96 | G9v3-39A5BAI9Stars | 52 | — | — | — | 128k | |
| 97 | Gemini 3.1 Pro PreviewGoogle | 52 | $4.50 | 134 | 24s | 1M | |
| 98 | Grok 4.20 0309 v2xAI | 52 | $1.56 | 120 | 19s | 2M | |
| 99 | MiMo-V2-OmniXiaomi | 52 | — | — | — | 256k | |
| 100 | Gemini 3.5 Flash-LiteGoogle | 51 | $0.85 | 419 | 8.3s | 1M | |
| 101 | GLM 5V TurboZ AI | 51 | — | — | — | 200k | |
| 102 | Ling-3.0-flash-FinInclusionAI | 51 | — | — | — | 256k | |
| 103 | K2 Horizon 7BInstitute of Foundation Models | 50 | — | — | — | 512k | |
| 104 | Kimi K2.5Kimi | 50 | $1.20 | 83 | 1.7s | 256k | |
| 105 | Qwen3.7 PlusAlibaba | 50 | $0.70 | 58 | 2.3s | 1M | |
| 106 | GPT-5.5 Instant (May 2026)OpenAI | 49 | $11.25 | — | — | 400k | |
| 107 | Hy3-previewTencent | 49 | $0.10 | — | — | 256k | Frontier |
| 108 | Ling 3.0 FlashInclusionAI | 49 | $0.11 | 245 | 2.6s | 256k | |
| 109 | MiniMax-M2.5MiniMax | 49 | $0.53 | 58 | 1.7s | 200k | |
| 110 | Nemotron 3.5 Bolt 74B-A9BNVIDIA | 49 | — | — | — | 256k | |
| 111 | A.X-K2SK Telecom | 48 | — | — | — | 256k | |
| 112 | Grok 4.3xAI | 48 | $1.56 | 102 | 27s | 1M | |
| 113 | MiniMax-M2.7MiniMax | 48 | $0.53 | 57 | 1.7s | 200k | |
| 114 | Qwen3 Max ThinkingAlibaba | 48 | — | — | — | 256k | |
| 115 | Qwen3.5 27BAlibaba | 48 | $0.83 | 78 | 5.7s | 256k | |
| 116 | Qwen3.5 Omni PlusAlibaba | 48 | $1.50 | 120 | 2.1s | 256k | |
| 117 | GPT-5.4 miniOpenAI | 47 | $1.69 | 237 | 107s | 400k | |
| 118 | Qwen3.5 35B A3BAlibaba | 47 | $0.69 | 151 | 2.1s | 256k | |
| 119 | KAT Coder Pro V2KwaiKAT | 46 | — | — | — | 256k | |
| 120 | MiMo-V2.5Xiaomi | 46 | $0.17 | 39 | 2.0s | 1M | |
| 121 | Qwen3.6 27BAlibaba | 46 | $1.35 | 57 | 3.6s | 256k | |
| 122 | Step 3.7 FlashStepFun | 46 | $0.44 | 113 | 2.5s | 256k | |
| 123 | K-EXAONE 2.0 0803LG AI Research | 45 | — | — | — | 256k | |
| 124 | JT-35B-FlashChina Mobile | 44 | — | — | — | 256k | |
| 125 | LongCat 2.0LongCat | 44 | $1.30 | 72 | 2.5s | 1M | |
| 126 | Qwen3.6 35B A3BAlibaba | 44 | $0.84 | 132 | 2.1s | 256k | |
| 127 | K2 Horizon 3.7BInstitute of Foundation Models | 42 | — | — | — | 512k | |
| 128 | Muse GlimmerMeta | 41 | $0.53 | 105 | 1.1s | 128k | |
| 129 | GPT-5.4 nanoOpenAI | 40 | $0.46 | 170 | 103s | 400k | |
| 130 | Ling-2.6-1TInclusionAI | 40 | $0.85 | — | — | 256k | |
| 131 | Step 3.5 Flash 2603StepFun | 40 | $0.15 | 150 | 3.4s | 256k | |
| 132 | Nemotron 3.5 LightningNVIDIA | 39 | $0.10 | 308 | 0.6s | 1M | Frontier |
| 133 | Qwen3.5 397B A17BAlibaba | 39 | $1.35 | 59 | 2.3s | 256k | |
| 134 | Step 3.5 FlashStepFun | 39 | $0.15 | 169 | 3.4s | 256k | |
| 135 | Gemini 3.1 Flash-LiteGoogle | 38 | $0.56 | 362 | 4.7s | 1M | |
| 136 | Gemma 4 26B A4BGoogle | 38 | $0.16 | 95 | 0.9s | 256k | |
| 137 | Qwen3.5 122B A10BAlibaba | 38 | $1.10 | 132 | 2.3s | 256k | |
| 138 | Ring-2.6-1TInclusionAI | 38 | $0.85 | 122 | 3.7s | 256k | |
| 139 | Mistral Medium 3.5Mistral | 37 | $3.00 | 163 | 2.3s | 256k | |
| 140 | Command A+Cohere | 36 | — | 160 | 0.5s | 192k | |
| 141 | Mercury 2.5Inception | 36 | $0.38 | 714 | 3.5s | 260k | |
| 142 | GLM-4.7-FlashZ AI | 35 | $0.15 | 78 | 1.3s | 200k | |
| 143 | MiniCPM5-2BOpenBMB | 35 | — | — | — | 128k | |
| 144 | Celeris-1Celeris | 34 | $0.33 | 1521 | 0.6s | 128k | |
| 145 | Nemotron 3 Super 120B A12BNVIDIA | 34 | $0.45 | 169 | 1.3s | 1M | |
| 146 | Qwen3.5 Omni FlashAlibaba | 34 | $0.28 | 235 | 1.8s | 256k | |
| 147 | Trinity Large ThinkingArcee AI | 34 | $0.39 | 347 | 1.2s | 512k | |
| 148 | EXAONE 4.5 33BLG AI Research | 33 | — | — | — | 256k | |
| 149 | Gemma 4 12BGoogle | 33 | $0.15 | 113 | 2.2s | 256k | |
| 150 | JT-MINIChina Mobile | 33 | — | — | — | 128k | |
| 151 | Mistral Small 4Mistral | 33 | $0.26 | 175 | 0.8s | 256k | |
| 152 | Nemotron 3 Nano Omni 30B A3B ReasoningNVIDIA | 33 | $0.64 | 232 | 0.7s | 256k | |
| 153 | Gemma 4 31BGoogle | 32 | — | 35 | 1.1s | 256k | |
| 154 | HyperNova 60B 2605Multiverse Computing | 32 | — | — | — | 128k | |
| 155 | LongCat Flash LiteLongCat | 32 | — | — | — | 256k | |
| 156 | Mercury 2Inception | 32 | $0.38 | 508 | 5.6s | 128k | |
| 157 | Nemotron Cascade 2 30B A3BNVIDIA | 32 | — | — | — | 1M | |
| 158 | G9v3-3BAI9Stars | 31 | — | — | — | 128k | |
| 159 | Granite 4.2 30BIBM | 31 | $0.28 | 39 | 0.9s | 128k | |
| 160 | LFM2.5-2.6BLiquid AI | 30 | — | — | — | 128k | |
| 161 | Nanbeige4.1-3BNanbeige | 30 | — | — | — | 256k | |
| 162 | Sarvam 105BSarvam | 30 | $0.07 | 195 | 1.6s | 128k | Frontier |
| 163 | DiffusionGemma 26B A4BGoogle | 29 | — | — | — | 256k | |
| 164 | Falcon-H1R-7BTII UAE | 29 | — | — | — | 256k | |
| 165 | Gemma 4 E2BGoogle | 29 | — | — | — | 128k | |
| 166 | Granite 4.1 30BIBM | 29 | — | — | — | 128k | |
| 167 | Granite 4.2 8BIBM | 29 | $0.11 | 40 | 0.9s | 128k | |
| 168 | Ling 2.6 FlashInclusionAI | 29 | — | — | — | 256k | |
| 169 | MiniCPM5-1BOpenBMB | 29 | — | — | — | 128k | |
| 170 | North Mini CodeCohere | 29 | — | 65 | 0.5s | 256k | |
| 171 | Qwen3.5 4BAlibaba | 29 | $0.06 | 20 | 0.9s | 256k | Frontier |
| 172 | Step3 VL 10BStepFun | 29 | — | — | — | 64k | |
| 173 | Tri-21B-ThinkTrillion Labs | 29 | — | — | — | 32k | |
| 174 | Tri-21B-think PreviewTrillion Labs | 29 | — | — | — | 32k | |
| 175 | Granite 4.1 8BIBM | 28 | — | — | — | 128k | |
| 176 | LFM2.5-8B-A1BLiquid AI | 28 | — | — | — | 32k | |
| 177 | Ling 3.0 TinyInclusionAI | 28 | — | 57 | 2.7s | 256k | |
| 178 | Olmo 3.1 32B InstructAllen Institute for AI | 28 | — | — | — | 64k | |
| 179 | Qwen3 Coder NextAlibaba | 28 | $0.48 | 96 | 0.9s | 256k | |
| 180 | Qwen3.5 2BAlibaba | 28 | — | — | — | 256k | |
| 181 | Sarvam 30BSarvam | 28 | — | — | — | 64k | |
| 182 | Solar Pro 3Upstage | 28 | $0.26 | 146 | 1.3s | 128k | |
| 183 | Granite 4.1 3BIBM | 27 | — | — | — | 128k | |
| 184 | Granite 4.2 3BIBM | 27 | $0.05 | 40 | 0.7s | 128k | Frontier |
| 185 | LFM2 24B A2BLiquid AI | 27 | — | — | — | 32k | |
| 186 | LFM2.5-1.2B-InstructLiquid AI | 27 | — | — | — | 32k | |
| 187 | LFM2.5-1.2B-ThinkingLiquid AI | 27 | — | — | — | 32k | |
| 188 | LFM2.5-VL-1.6BLiquid AI | 27 | — | — | — | 32k | |
| 189 | MiniCPM-V 4.6 1.3BOpenBMB | 27 | — | — | — | 262k | |
| 190 | Qwen3.5 0.8BAlibaba | 27 | — | — | — | 256k | |
| 191 | Qwen3.5 9BAlibaba | 27 | $0.15 | 60 | 2.0s | 256k | |
| 192 | Tiny Aya GlobalCohere | 27 | — | — | — | 8k | |
| 193 | Gemma 4 E4BGoogle | 26 | $0.04 | 46 | 1.1s | 128k | Frontier |
| 194 | NVIDIA Nemotron 3 Nano 4BNVIDIA | 25 | — | — | — | 262k | |
| 195 | K2 Horizon 0.9BInstitute of Foundation Models | 23 | — | — | — | 128k |
Muse Wonder Full is included in Muse Code Power Usage ($50/month) and Maximum Usage ($200/month). API token pricing and throughput are unavailable.
Mistral API public preview: Mistral launched October 6 with 1M context and a two-week offer: $0.68 input, $0.07 cached input, and $2.09 output per 1M tokens. Tables retain Artificial Analysis’s measured pricing and context.
Compare
- Kimi K3.1 ProKimi100
- Claude Opus 5.5Anthropic98
- Claude Sonnet 5.5Anthropic96
- Muse WonderMeta Model API and Muse Code96
- Muse Wonder FullMuse Code subscription96
- DeepSeek V4.1 Flash 1017DeepSeek93
- Claude Fable 5.1Anthropic92
- Gemini 4 ArgonGoogle91
- Claude Opus 5Anthropic90
- Muse Spark 1.3Meta89
- Step 5 PreviewStepFun89
- GPT-6 AstraOpenAI88
- Claude Fable 5Anthropic87
- GLM-5.3Z AI87
- GPT-6.1 SolOpenAI87
- Grok 4.7xAI87
- MiMo-V2.6-ProXiaomi87
- MiniMax-M3.1-FlashMiniMax87
- Qwen3.8 Max (0902)Alibaba86
- DeepSeek V4.1 FlashDeepSeek85
- Ling 3.1 FlashInclusionAI85
- Ling 3.1 Flash 1019InclusionAI85
- Grok 4.6xAI84
- Kimi K3Kimi84
- GPT-5.6 SolOpenAI83
- GPT-6 SolOpenAI83
- GLM-5.3-FlashZ AI82
- Muse Spark OpenMeta82
- GPT-6.1 TerraOpenAI81
- Muse Spark 1.2Meta81
- Qwen3.8-Flash-NextAlibaba81
- Claude Haiku 5.5Anthropic79
- Qwen3.8 2.4T A95BAlibaba79
- DeepSeek V4 Flash VisionDeepSeek78
- Mistral Large 4 PreviewMistral API public preview78
- Qwen3.8 MaxAlibaba78
- Gemini 3.8 FlashGoogle77
- MiMo-V2.6-FlashXiaomi77
- Claude Opus 4.7Anthropic76
- DeepSeek V4 Pro 0813DeepSeek76
- Gemini 3.7 FlashGoogle76
- Agnes 2.5 Pro BetaSapiens AI75
- Claude Opus 4.8Anthropic75
- DeepSeek V4 Flash 0731DeepSeek75
- GPT-5.6 TerraOpenAI75
- Grok 4.5xAI74
- K2 Horizon 375B A23BInstitute of Foundation Models74
- GPT-6 LunaOpenAI71
- Beam 1.0Reflection70
- GPT-5.6 LunaOpenAI70
- JT-4.1 Flash 236B A21BChina Mobile70
- Claude Sonnet 5Anthropic69
- GLM-5.2Z AI69
- GPT-5.4OpenAI69
- GPT-5.5OpenAI69
- Qwen3.8 27BAlibaba69
- Gemini 3.6 FlashGoogle68
- Motif 3Motif Technologies68
- Muse Spark 1.1Meta68
- Motif 3 (Beta)Motif Technologies66
- Gemini 3.5 FlashGoogle65
- Nemotron 3.5 Super VLNVIDIA65
- MiniMax-M3MiniMax64
- DeepSeek V4 Pro 0424DeepSeek63
- Qwen3.7 MaxAlibaba63
- Muse SparkMeta61
- Quasar 438BMultiverse Computing61
- Apodex 1.1Apodex60
- Solar Pro 4Upstage60
- Claude Opus 4.6Anthropic59
- GLM-5Z AI59
- Grok Build 0.1 0616xAI59
- Inkling SmallThinking Machines59
- MiMo-V2-ProXiaomi59
- GLM-5-TurboZ AI58
- K2 Horizon MoVA 36B A4BInstitute of Foundation Models58
- Nex-N2-ProNex AGI58
- Qwen3.6 Max PreviewAlibaba58
- GLM-5.1Z AI57
- GPT-5.3 CodexOpenAI57
- InklingThinking Machines57
- Claude Sonnet 4.6Anthropic55
- GPT-5.5 Instant (June 2026)OpenAI55
- Grok 4.20 0309xAI55
- Kimi K2.6Kimi55
- Kimi K2.7 CodeKimi55
- MiMo-V2-Omni-0327Xiaomi55
- MiMo-V2.5-ProXiaomi55
- Nemotron 3 Ultra 550B A55BNVIDIA55
- Qwen3.6 PlusAlibaba55
- Hy3Tencent54
- Ling-3.0-flash-VLInclusionAI54
- Solar Mini 4Upstage54
- DeepSeek V4 Flash 0420DeepSeek53
- Solar Open2 250BUpstage53
- G9v3-39A5BAI9Stars52
- Gemini 3.1 Pro PreviewGoogle52
- Grok 4.20 0309 v2xAI52
- MiMo-V2-OmniXiaomi52
- Gemini 3.5 Flash-LiteGoogle51
- GLM 5V TurboZ AI51
- Ling-3.0-flash-FinInclusionAI51
- K2 Horizon 7BInstitute of Foundation Models50
- Kimi K2.5Kimi50
- Qwen3.7 PlusAlibaba50
- GPT-5.5 Instant (May 2026)OpenAI49
- Hy3-previewTencent49
- Ling 3.0 FlashInclusionAI49
- MiniMax-M2.5MiniMax49
- Nemotron 3.5 Bolt 74B-A9BNVIDIA49
- A.X-K2SK Telecom48
- Grok 4.3xAI48
- MiniMax-M2.7MiniMax48
- Qwen3 Max ThinkingAlibaba48
- Qwen3.5 27BAlibaba48
- Qwen3.5 Omni PlusAlibaba48
- GPT-5.4 miniOpenAI47
- Qwen3.5 35B A3BAlibaba47
- KAT Coder Pro V2KwaiKAT46
- MiMo-V2.5Xiaomi46
- Qwen3.6 27BAlibaba46
- Step 3.7 FlashStepFun46
- K-EXAONE 2.0 0803LG AI Research45
- JT-35B-FlashChina Mobile44
- LongCat 2.0LongCat44
- Qwen3.6 35B A3BAlibaba44
- K2 Horizon 3.7BInstitute of Foundation Models42
- Muse GlimmerMeta41
- GPT-5.4 nanoOpenAI40
- Ling-2.6-1TInclusionAI40
- Step 3.5 Flash 2603StepFun40
- Nemotron 3.5 LightningNVIDIA39
- Qwen3.5 397B A17BAlibaba39
- Step 3.5 FlashStepFun39
- Gemini 3.1 Flash-LiteGoogle38
- Gemma 4 26B A4BGoogle38
- Qwen3.5 122B A10BAlibaba38
- Ring-2.6-1TInclusionAI38
- Mistral Medium 3.5Mistral37
- Command A+Cohere36
- Mercury 2.5Inception36
- GLM-4.7-FlashZ AI35
- MiniCPM5-2BOpenBMB35
- Celeris-1Celeris34
- Nemotron 3 Super 120B A12BNVIDIA34
- Qwen3.5 Omni FlashAlibaba34
- Trinity Large ThinkingArcee AI34
- EXAONE 4.5 33BLG AI Research33
- Gemma 4 12BGoogle33
- JT-MINIChina Mobile33
- Mistral Small 4Mistral33
- Nemotron 3 Nano Omni 30B A3B ReasoningNVIDIA33
- Gemma 4 31BGoogle32
- HyperNova 60B 2605Multiverse Computing32
- LongCat Flash LiteLongCat32
- Mercury 2Inception32
- Nemotron Cascade 2 30B A3BNVIDIA32
- G9v3-3BAI9Stars31
- Granite 4.2 30BIBM31
- LFM2.5-2.6BLiquid AI30
- Nanbeige4.1-3BNanbeige30
- Sarvam 105BSarvam30
- DiffusionGemma 26B A4BGoogle29
- Falcon-H1R-7BTII UAE29
- Gemma 4 E2BGoogle29
- Granite 4.1 30BIBM29
- Granite 4.2 8BIBM29
- Ling 2.6 FlashInclusionAI29
- MiniCPM5-1BOpenBMB29
- North Mini CodeCohere29
- Qwen3.5 4BAlibaba29
- Step3 VL 10BStepFun29
- Tri-21B-ThinkTrillion Labs29
- Tri-21B-think PreviewTrillion Labs29
- Granite 4.1 8BIBM28
- LFM2.5-8B-A1BLiquid AI28
- Ling 3.0 TinyInclusionAI28
- Olmo 3.1 32B InstructAllen Institute for AI28
- Qwen3 Coder NextAlibaba28
- Qwen3.5 2BAlibaba28
- Sarvam 30BSarvam28
- Solar Pro 3Upstage28
- Granite 4.1 3BIBM27
- Granite 4.2 3BIBM27
- LFM2 24B A2BLiquid AI27
- LFM2.5-1.2B-InstructLiquid AI27
- LFM2.5-1.2B-ThinkingLiquid AI27
- LFM2.5-VL-1.6BLiquid AI27
- MiniCPM-V 4.6 1.3BOpenBMB27
- Qwen3.5 0.8BAlibaba27
- Qwen3.5 9BAlibaba27
- Tiny Aya GlobalCohere27
- Gemma 4 E4BGoogle26
- NVIDIA Nemotron 3 Nano 4BNVIDIA25
- K2 Horizon 0.9BInstitute of Foundation Models23
- Kimi K3.1 ProKimi100
- Claude Opus 5.5Anthropic98
- Claude Sonnet 5.5Anthropic96
- Muse WonderMeta Model API and Muse Code96
- Muse Wonder FullMuse Code subscription96
- DeepSeek V4.1 Flash 1017DeepSeek93
- Claude Fable 5.1Anthropic92
- Gemini 4 ArgonGoogle91
- Claude Opus 5Anthropic90
- Muse Spark 1.3Meta89
- Step 5 PreviewStepFun89
- GPT-6 AstraOpenAI88
- Claude Fable 5Anthropic87
- GLM-5.3Z AI87
- GPT-6.1 SolOpenAI87
- Grok 4.7xAI87
- MiMo-V2.6-ProXiaomi87
- MiniMax-M3.1-FlashMiniMax87
- Qwen3.8 Max (0902)Alibaba86
- DeepSeek V4.1 FlashDeepSeek85
- Ling 3.1 FlashInclusionAI85
- Ling 3.1 Flash 1019InclusionAI85
- Grok 4.6xAI84
- Kimi K3Kimi84
- GPT-5.6 SolOpenAI83
- GPT-6 SolOpenAI83
- GLM-5.3-FlashZ AI82
- Muse Spark OpenMeta82
- GPT-6.1 TerraOpenAI81
- Muse Spark 1.2Meta81
- Qwen3.8-Flash-NextAlibaba81
- Claude Haiku 5.5Anthropic79
- Qwen3.8 2.4T A95BAlibaba79
- DeepSeek V4 Flash VisionDeepSeek78
- Mistral Large 4 PreviewMistral API public preview78
- Qwen3.8 MaxAlibaba78
- Gemini 3.8 FlashGoogle77
- MiMo-V2.6-FlashXiaomi77
- Claude Opus 4.7Anthropic76
- DeepSeek V4 Pro 0813DeepSeek76
- Gemini 3.7 FlashGoogle76
- Agnes 2.5 Pro BetaSapiens AI75
- Claude Opus 4.8Anthropic75
- DeepSeek V4 Flash 0731DeepSeek75
- GPT-5.6 TerraOpenAI75
- Grok 4.5xAI74
- K2 Horizon 375B A23BInstitute of Foundation Models74
- GPT-6 LunaOpenAI71
- Beam 1.0Reflection70
- GPT-5.6 LunaOpenAI70
- JT-4.1 Flash 236B A21BChina Mobile70
- Claude Sonnet 5Anthropic69
- GLM-5.2Z AI69
- GPT-5.4OpenAI69
- GPT-5.5OpenAI69
- Qwen3.8 27BAlibaba69
- Gemini 3.6 FlashGoogle68
- Motif 3Motif Technologies68
- Muse Spark 1.1Meta68
- Motif 3 (Beta)Motif Technologies66
- Gemini 3.5 FlashGoogle65
- Nemotron 3.5 Super VLNVIDIA65
- MiniMax-M3MiniMax64
- DeepSeek V4 Pro 0424DeepSeek63
- Qwen3.7 MaxAlibaba63
- Muse SparkMeta61
- Quasar 438BMultiverse Computing61
- Apodex 1.1Apodex60
- Solar Pro 4Upstage60
- Claude Opus 4.6Anthropic59
- GLM-5Z AI59
- Grok Build 0.1 0616xAI59
- Inkling SmallThinking Machines59
- MiMo-V2-ProXiaomi59
- GLM-5-TurboZ AI58
- K2 Horizon MoVA 36B A4BInstitute of Foundation Models58
- Nex-N2-ProNex AGI58
- Qwen3.6 Max PreviewAlibaba58
- GLM-5.1Z AI57
- GPT-5.3 CodexOpenAI57
- InklingThinking Machines57
- Claude Sonnet 4.6Anthropic55
- GPT-5.5 Instant (June 2026)OpenAI55
- Grok 4.20 0309xAI55
- Kimi K2.6Kimi55
- Kimi K2.7 CodeKimi55
- MiMo-V2-Omni-0327Xiaomi55
- MiMo-V2.5-ProXiaomi55
- Nemotron 3 Ultra 550B A55BNVIDIA55
- Qwen3.6 PlusAlibaba55
- Hy3Tencent54
- Ling-3.0-flash-VLInclusionAI54
- Solar Mini 4Upstage54
- DeepSeek V4 Flash 0420DeepSeek53
- Solar Open2 250BUpstage53
- G9v3-39A5BAI9Stars52
- Gemini 3.1 Pro PreviewGoogle52
- Grok 4.20 0309 v2xAI52
- MiMo-V2-OmniXiaomi52
- Gemini 3.5 Flash-LiteGoogle51
- GLM 5V TurboZ AI51
- Ling-3.0-flash-FinInclusionAI51
- K2 Horizon 7BInstitute of Foundation Models50
- Kimi K2.5Kimi50
- Qwen3.7 PlusAlibaba50
- GPT-5.5 Instant (May 2026)OpenAI49
- Hy3-previewTencent49
- Ling 3.0 FlashInclusionAI49
- MiniMax-M2.5MiniMax49
- Nemotron 3.5 Bolt 74B-A9BNVIDIA49
- A.X-K2SK Telecom48
- Grok 4.3xAI48
- MiniMax-M2.7MiniMax48
- Qwen3 Max ThinkingAlibaba48
- Qwen3.5 27BAlibaba48
- Qwen3.5 Omni PlusAlibaba48
- GPT-5.4 miniOpenAI47
- Qwen3.5 35B A3BAlibaba47
- KAT Coder Pro V2KwaiKAT46
- MiMo-V2.5Xiaomi46
- Qwen3.6 27BAlibaba46
- Step 3.7 FlashStepFun46
- K-EXAONE 2.0 0803LG AI Research45
- JT-35B-FlashChina Mobile44
- LongCat 2.0LongCat44
- Qwen3.6 35B A3BAlibaba44
- K2 Horizon 3.7BInstitute of Foundation Models42
- Muse GlimmerMeta41
- GPT-5.4 nanoOpenAI40
- Ling-2.6-1TInclusionAI40
- Step 3.5 Flash 2603StepFun40
- Nemotron 3.5 LightningNVIDIA39
- Qwen3.5 397B A17BAlibaba39
- Step 3.5 FlashStepFun39
- Gemini 3.1 Flash-LiteGoogle38
- Gemma 4 26B A4BGoogle38
- Qwen3.5 122B A10BAlibaba38
- Ring-2.6-1TInclusionAI38
- Mistral Medium 3.5Mistral37
- Command A+Cohere36
- Mercury 2.5Inception36
- GLM-4.7-FlashZ AI35
- MiniCPM5-2BOpenBMB35
- Celeris-1Celeris34
- Nemotron 3 Super 120B A12BNVIDIA34
- Qwen3.5 Omni FlashAlibaba34
- Trinity Large ThinkingArcee AI34
- EXAONE 4.5 33BLG AI Research33
- Gemma 4 12BGoogle33
- JT-MINIChina Mobile33
- Mistral Small 4Mistral33
- Nemotron 3 Nano Omni 30B A3B ReasoningNVIDIA33
- Gemma 4 31BGoogle32
- HyperNova 60B 2605Multiverse Computing32
- LongCat Flash LiteLongCat32
- Mercury 2Inception32
- Nemotron Cascade 2 30B A3BNVIDIA32
- G9v3-3BAI9Stars31
- Granite 4.2 30BIBM31
- LFM2.5-2.6BLiquid AI30
- Nanbeige4.1-3BNanbeige30
- Sarvam 105BSarvam30
- DiffusionGemma 26B A4BGoogle29
- Falcon-H1R-7BTII UAE29
- Gemma 4 E2BGoogle29
- Granite 4.1 30BIBM29
- Granite 4.2 8BIBM29
- Ling 2.6 FlashInclusionAI29
- MiniCPM5-1BOpenBMB29
- North Mini CodeCohere29
- Qwen3.5 4BAlibaba29
- Step3 VL 10BStepFun29
- Tri-21B-ThinkTrillion Labs29
- Tri-21B-think PreviewTrillion Labs29
- Granite 4.1 8BIBM28
- LFM2.5-8B-A1BLiquid AI28
- Ling 3.0 TinyInclusionAI28
- Olmo 3.1 32B InstructAllen Institute for AI28
- Qwen3 Coder NextAlibaba28
- Qwen3.5 2BAlibaba28
- Sarvam 30BSarvam28
- Solar Pro 3Upstage28
- Granite 4.1 3BIBM27
- Granite 4.2 3BIBM27
- LFM2 24B A2BLiquid AI27
- LFM2.5-1.2B-InstructLiquid AI27
- LFM2.5-1.2B-ThinkingLiquid AI27
- LFM2.5-VL-1.6BLiquid AI27
- MiniCPM-V 4.6 1.3BOpenBMB27
- Qwen3.5 0.8BAlibaba27
- Qwen3.5 9BAlibaba27
- Tiny Aya GlobalCohere27
- Gemma 4 E4BGoogle26
- NVIDIA Nemotron 3 Nano 4BNVIDIA25
- K2 Horizon 0.9BInstitute of Foundation Models23
100Klu Index98
$4 / 1M · 46 tok/sPrice and speed$8 / 1M · 98 tok/s
Frontier benchmark scores
| Model | T-B Science | Chartography | ProgramBench | T-B 4.0 | Harvey's Lab | HealthBench Professional | CursorBench | FrontierCode |
|---|---|---|---|---|---|---|---|---|
| Kimi K3.1 Pro | Not reported | Not reported | Not reported | Not reported | Not reported | Not reported | Not reported | Not reported |
| Claude Opus 5.5 | 47.1 | 66.3 | 18.5 | 65.2 | 3.8 | 59.9 | 57.8 | 54.6 |
| Claude Sonnet 5.5 | 45.7 | Not reported | 6.5 | 64.1 | Not reported | 63.0 | 55.5 | 52.1 |
| Muse Wonder | Not reported | Not reported | Not reported | Not reported | Not reported | Not reported | Not reported | Not reported |
| Muse Wonder Full | Not reported | Not reported | Not reported | Not reported | Not reported | Not reported | Not reported | Not reported |
| DeepSeek V4.1 Flash 1017 | 0.0 | 11.6 | Not reported | Not reported | Not reported | 50.7 | Not reported | Not reported |
| Claude Fable 5.1 | 40.0 | 46.2 | 7.0 | 58.1 | 6.7 | 57.8 | 51.8 | 50.9 |
| Gemini 4 Argon | 44.3 | 71.6 | 2.5 | 57.6 | 19.6 | 55.1 | Not reported | Not reported |
| Claude Opus 5 | 27.1 | 27.3 | 3.0 | 53.5 | 6.7 | Not reported | 46.6 | 53.4 |
| Muse Spark 1.3 | 10.0 | 27.6 | 2.5 | Not reported | 23.8 | Not reported | 41.6 | Not reported |
| Step 5 Preview | 2.9 | Not reported | Not reported | 31.8 | Not reported | Not reported | Not reported | Not reported |
| GPT-6 Astra | 62.9 | 71.0 | 5.5 | 59.6 | 5.4 | 66.0 | Not reported | 53.3 |
How the Klu Index works
The Klu Index combines accuracy, use-case evaluations, human preference, and speed into one score out of 100. The fastest way to compare models.
AccuracyPublic benchmark results across reasoning, coding, and agentic work35%
EvaluationsLegal, finance, healthcare, engineering, economics, and strategy30%
PreferenceBlind pairwise judgments of finished work20%
PerformanceOutput speed and time to first token15%
Models by lab
Kimi5 models
ModelKlu Index$ / 1M
Kimi K3.1 Pro100$4.00
Kimi K384$6.00
Kimi K2.655$1.71
Kimi K2.7 Code55$1.71
Kimi K2.550$1.20
Anthropic11 models
ModelKlu Index$ / 1M
Claude Opus 5.598$8.00
Claude Sonnet 5.596$4.00
Claude Fable 5.192$20.00
Claude Opus 590$10.00
Claude Fable 587$20.00
Claude Haiku 5.579$0.20
Claude Opus 4.776$10.00
Claude Opus 4.875$10.00
Claude Sonnet 569$4.00
Claude Opus 4.659$10.00
Claude Sonnet 4.655$6.00
Meta8 models
ModelKlu Index$ / 1M
Muse Wonder96$8.00
Muse Wonder FullMuse Code subscription96—
Muse Spark 1.389$2.00
Muse Spark Open82$2.00
Muse Spark 1.281$2.00
Muse Spark 1.168$2.00
Muse Spark61—
Muse Glimmer41$0.53
DeepSeek7 models
ModelKlu Index$ / 1M
DeepSeek V4.1 Flash 101793$0.53
DeepSeek V4.1 Flash85$0.52
DeepSeek V4 Flash Vision78$0.66
DeepSeek V4 Pro 081376$1.98
DeepSeek V4 Flash 073175$0.18
DeepSeek V4 Pro 042463$0.54
DeepSeek V4 Flash 042053$0.14
Google14 models
ModelKlu Index$ / 1M
Gemini 4 Argon91$4.00
Gemini 3.8 Flash77$1.50
Gemini 3.7 Flash76$1.50
Gemini 3.6 Flash68$1.50
Gemini 3.5 Flash65$3.38
Gemini 3.1 Pro Preview52$4.50
Gemini 3.5 Flash-Lite51$0.85
Gemini 3.1 Flash-Lite38$0.56
Gemma 4 26B A4B38$0.16
Gemma 4 12B33$0.15
Gemma 4 31B32—
DiffusionGemma 26B A4B29—
Gemma 4 E2B29—
Gemma 4 E4B26$0.04
StepFun5 models
ModelKlu Index$ / 1M
Step 5 Preview89$1.43
Step 3.7 Flash46$0.44
Step 3.5 Flash 260340$0.15
Step 3.5 Flash39$0.15
Step3 VL 10B29—
OpenAI15 models
ModelKlu Index$ / 1M
GPT-6 Astra88$20.00
GPT-6.1 Sol87$4.00
GPT-5.6 Sol83$8.00
GPT-6 Sol83$4.00
GPT-6.1 Terra81$2.00
GPT-5.6 Terra75$4.50
GPT-6 Luna71$0.20
GPT-5.6 Luna70$0.45
GPT-5.469$5.63
GPT-5.569$11.25
GPT-5.3 Codex57$4.81
GPT-5.5 Instant (June 2026)55$11.25
GPT-5.5 Instant (May 2026)49$11.25
GPT-5.4 mini47$1.69
GPT-5.4 nano40$0.46
Z AI8 models
ModelKlu Index$ / 1M
GLM-5.387$2.15
GLM-5.3-Flash82$0.24
GLM-5.269$2.15
GLM-559$1.55
GLM-5-Turbo58—
GLM-5.157$1.98
GLM 5V Turbo51—
GLM-4.7-Flash35$0.15
xAI7 models
ModelKlu Index$ / 1M
Grok 4.787$3.00
Grok 4.684$3.00
Grok 4.574$3.00
Grok Build 0.1 061659$1.25
Grok 4.20 030955$3.00
Grok 4.20 0309 v252$1.56
Grok 4.348$1.56
Xiaomi7 models
ModelKlu Index$ / 1M
MiMo-V2.6-Pro87$0.54
MiMo-V2.6-Flash77$0.18
MiMo-V2-Pro59—
MiMo-V2-Omni-032755—
MiMo-V2.5-Pro55$0.56
MiMo-V2-Omni52—
MiMo-V2.546$0.17
MiniMax4 models
ModelKlu Index$ / 1M
MiniMax-M3.1-Flash87$0.26
MiniMax-M364$0.53
MiniMax-M2.549$0.53
MiniMax-M2.748$0.53
Alibaba23 models
ModelKlu Index$ / 1M
Qwen3.8 Max (0902)86$3.00
Qwen3.8-Flash-Next81$0.23
Qwen3.8 2.4T A95B79$3.00
Qwen3.8 Max78$3.00
Qwen3.8 27B69$1.13
Qwen3.7 Max63$3.75
Qwen3.6 Max Preview58—
Qwen3.6 Plus55$1.13
Qwen3.7 Plus50$0.70
Qwen3 Max Thinking48—
Qwen3.5 27B48$0.83
Qwen3.5 Omni Plus48$1.50
Qwen3.5 35B A3B47$0.69
Qwen3.6 27B46$1.35
Qwen3.6 35B A3B44$0.84
Qwen3.5 397B A17B39$1.35
Qwen3.5 122B A10B38$1.10
Qwen3.5 Omni Flash34$0.28
Qwen3.5 4B29$0.06
Qwen3 Coder Next28$0.48
Qwen3.5 2B28—
Qwen3.5 0.8B27—
Qwen3.5 9B27$0.15
InclusionAI9 models
ModelKlu Index$ / 1M
Ling 3.1 Flash85$0.45
Ling 3.1 Flash 101985$0.45
Ling-3.0-flash-VL54—
Ling-3.0-flash-Fin51—
Ling 3.0 Flash49$0.11
Ling-2.6-1T40$0.85
Ring-2.6-1T38$0.85
Ling 2.6 Flash29—
Ling 3.0 Tiny28—
Mistral3 models
ModelKlu Index$ / 1M
Mistral Large 4 Preview78$2.07
Mistral Medium 3.537$3.00
Mistral Small 433$0.26
Institute of Foundation Models5 models
ModelKlu Index$ / 1M
K2 Horizon 375B A23B74—
K2 Horizon MoVA 36B A4B58—
K2 Horizon 7B50—
K2 Horizon 3.7B42—
K2 Horizon 0.9B23—
China Mobile3 models
ModelKlu Index$ / 1M
JT-4.1 Flash 236B A21B70—
JT-35B-Flash44—
JT-MINI33—
NVIDIA8 models
ModelKlu Index$ / 1M
Nemotron 3.5 Super VL65$0.31
Nemotron 3 Ultra 550B A55B55$1.13
Nemotron 3.5 Bolt 74B-A9B49—
Nemotron 3.5 Lightning39$0.10
Nemotron 3 Super 120B A12B34$0.45
Nemotron 3 Nano Omni 30B A3B Reasoning33$0.64
Nemotron Cascade 2 30B A3B32—
NVIDIA Nemotron 3 Nano 4B25—
Upstage4 models
ModelKlu Index$ / 1M
Solar Pro 460$0.53
Solar Mini 454$0.18
Solar Open2 250B53—
Solar Pro 328$0.26
Thinking Machines2 models
ModelKlu Index$ / 1M
Inkling Small59$0.53
Inkling57$1.76
Cohere3 models
ModelKlu Index$ / 1M
Command A+36—
North Mini Code29—
Tiny Aya Global27—
OpenBMB3 models
ModelKlu Index$ / 1M
MiniCPM5-2B35—
MiniCPM5-1B29—
MiniCPM-V 4.6 1.3B27—
IBM6 models
ModelKlu Index$ / 1M
Granite 4.2 30B31$0.28
Granite 4.1 30B29—
Granite 4.2 8B29$0.11
Granite 4.1 8B28—
Granite 4.1 3B27—
Granite 4.2 3B27$0.05
Liquid AI6 models
ModelKlu Index$ / 1M
LFM2.5-2.6B30—
LFM2.5-8B-A1B28—
LFM2 24B A2B27—
LFM2.5-1.2B-Instruct27—
LFM2.5-1.2B-Thinking27—
LFM2.5-VL-1.6B27—
Other labs29 models
ModelKlu Index$ / 1M
Agnes 2.5 Pro Beta75$0.15
Beam 1.070$0.60
Motif 368—
Motif 3 (Beta)66—
Quasar 438B61$0.17
Apodex 1.160$0.98
Nex-N2-Pro58—
Hy354$0.24
G9v3-39A5B52—
Hy3-preview49$0.10
A.X-K248—
KAT Coder Pro V246—
K-EXAONE 2.0 080345—
LongCat 2.044$1.30
Mercury 2.536$0.38
Celeris-134$0.33
Trinity Large Thinking34$0.39
EXAONE 4.5 33B33—
HyperNova 60B 260532—
LongCat Flash Lite32—
Mercury 232$0.38
G9v3-3B31—
Nanbeige4.1-3B30—
Sarvam 105B30$0.07
Falcon-H1R-7B29—
Tri-21B-Think29—
Tri-21B-think Preview29—
Olmo 3.1 32B Instruct28—
Sarvam 30B28—
About the leaderboard
- What is the Klu Index?
- One score out of 100 that blends public benchmark accuracy, use-case evaluations, blind preference judgments, and speed. Scores are scaled so the top model lands at 100.
- How is speed measured?
- Tokens per second is the median output speed on each provider's API. First token is the median time, in seconds, before the first token arrives.
- How often is it updated?
- Each time Klu refreshes the snapshot. The date at the top of the page shows when the current numbers were pulled.
- Which model is best for me?
- The one that wins on your own evals. Klu runs every model on your evals and shows which one comes out ahead for each use case.
The leaderboard is a starting point
The model at the top here isn't always the one that wins on your work. Klu runs every model on your evals and reruns them each time a new one ships.