Skip to content

Newest models

Most recent releases, with whatever independent scores exist so far.

Most recent releases, with whatever independent scores exist so far. Source: OpenRouter / Epoch AI. Variants (reasoning effort, batch) are folded into one row.

#ModelCapabilityReasoningCodingContext$/M
1MiMo-V2.6-FlashNEWXiaomi———1.05M$0.17
2MiMo-V2.6-ProNEWXiaomi———1.05M$0.54
3Ternary Bonsai 2 27BNEWPrismML———262K$0.18
4Schematron V2 SmallNEWInference.net———128K$0.095
5Schematron V2 TurboNEWInference.net———128K$0.060
6DeepSeek V4.1 FlashNEWbatchDeepSeek · +2 variants———1.05M$0.17
7Ling 3.0 Flash VLNEWinclusionAI (Ant Group)———262K$0.031
8Nex-N2.5-MiniNEWNex AGI———262K$0.044
9Nex-N2.5-ProNEWNex AGI———262K$0.12
10Granite 4.2 8BNEWIBM———131K$0.11
11Hy4 previewTencent———1.05M$1.25
12Tencent Hy4 previewunknownTencent—————
13GLM 5.3 FlashbatchZ.ai (Zhipu AI) · +1 variant———1.05M$0.095
14Qwen3.8 FlashAlibaba (Qwen)———1M$0.23
15DeepSeek V4 Flash Vision ExpDeepSeek———1.05M$0.32
16GLM-5.3-FlashunknownZ.ai (Zhipu AI)—————
17Hy-MT2-1.8BTencent———8K$0.077
18Hy-MT2-30B-A3BTencent———8K$0.13
19Hy-MT2-7BTencent———8K$0.13
20GLM 5.3batchZ.ai (Zhipu AI) · +1 variant———1.05M$0.84
21Qwen 3.8 27BAlibaba (Qwen)149.4————
22Qwen3.8 27BAlibaba (Qwen) · +1 variant———1M$1.14
23DeepSeek V4 Pro 0813DeepSeek · +3 variants155.491.7—1.05M$1.41
24Qwen3.8 2.4T A95BAlibaba (Qwen)———1.05M$3.00
25Nemotron 3.5 LightningNVIDIA———1M$0.085
26Muse Glimmer 30BMeta———131K$0.53
27DeepSeek V4 Flash 0731DeepSeek · +3 variants154.591.0—1.31M$0.093
28Ling 3.0 FlashinclusionAI (Ant Group)———262K$0.032
29Laguna S 2.1Poolside———1.05M$0.11
30LongCat 2.0Meituan———1.05M$0.53
31Kimi K3Moonshot AI · +3 variants157.793.168.51.05M$6.00
32Inkling SmallThinking Machines Lab · +5 variants150.1——524K$0.64
33InklingThinking Machines Lab · +1 variant148.6——524K$1.76
34Hy3Tencent———262K$0.23
35Laguna XS 2.1Poolside———262K$0.075
36GLM 5.2Z.ai (Zhipu AI)151.891.943.81.05M$1.39
37Kimi K2.7 CodeMoonshot AI150.087.930.5262K$1.32
38Nemotron 3 UltraNVIDIA146.285.4—262K$1.05
39Nemotron 3.5 Content SafetyNVIDIA———131K$0.20
40MiniMax M3MiniMax147.090.9—1.05M$0.53
41Step 3.7 FlashStepFun———262K$0.44
42Mistral Medium 3.5Mistral AI141.4——262K$3.00
43Laguna XS.2Poolside—————
44Qwen3.6 35B A3BAlibaba (Qwen)—84.8—262K$0.36
45DeepSeek V4 Pro 0423DeepSeek149.1——1.05M$1.19
46DeepSeek V4 Flash 0423DeepSeek146.1——1.05M$0.17
47DeepSeek v4unknownDeepSeek—————
48DeepSeek v4 FlashmaxDeepSeek—————
49DeepSeek v4 Prounknown thinkingDeepSeek · +1 variant—73.2———
50Qwen3.6 27BAlibaba (Qwen)146.585.9—262K$1.04
— means no published score from that source yet · click a model for every benchmark with its source
Best AI for coding →
Models ranked on DeepSWE
Best AI for reasoning →
Ranked on GPQA Diamond
Best AI for math →
Ranked on competition math
Best open-weight models →
Weights you can run yourself
Cheapest capable models →
Lowest blended API price
Best AI coding tools →
Apps & agents by popularity
Best AI image generators →
Apps & models
Best AI for research →
Answer engines & deep research