VECTOR WIREAI INTELLIGENCE
NVDA$1,847+3.2%MSFT$512+1.1%GOOGL$199-0.4%META$728+2.7%AMD$184-1.2%TSM$212+0.6%PLTR$98+4.1%AI IDX4,821+1.9%
PKT
SEEDRefresh Models Deals Regulatory Sources

GSM8K

100 models tracked

GSM8K (Grade School Math 8K) — grade-school math word problems; the standard accuracy benchmark.

#ModelVendorBest scoreRunsLast seen
1MiMo V2.5 ProXiaomi99.622026-08-23
2MiMo V2.5 Pro BaseXiaomi99.632026-06-05
3Kimi K2 InstructMoonshot97.312026-08-23
4o1OpenAI97.112026-08-23
5GPT-4.5 PreviewOpenAI9712026-08-23
6Claude 3.5 SonnetAnthropic96.932026-08-23
7Llama 3.1 Instruct 405BMeta96.852026-08-23
8DeepSeek-V3DeepSeek96.722026-07-02
9Gemma 3 27BGoogle95.912026-08-23
10Qwen2.5 Instruct 72BAlibaba95.832026-08-23
11GPT-4oOpenAI95.632026-07-02
12Gemini 2.0 Flash ExpGoogle95.422026-06-06
13Hy3 preview-BaseTencent95.422026-06-05
14Gemini 1.5 ProGoogle95.242026-08-23
15DeepSeek-V2.5DeepSeek95.112026-08-23
16Claude 3 OpusAnthropic9522026-08-23
17Nova Pro (Non-Reasoning)Amazon94.832026-08-23
18MiniMax Text 01MiniMax94.812026-06-06
19Gemini 2.0 Flash (experimental)Google94.612026-07-02
20Nova Lite (Non-Reasoning)Amazon94.512026-08-23
21Qwen3.5 122B A10BAlibaba94.512026-07-05
22Gemma 3 12BGoogle94.412026-08-23
23Qwen3 235B A22BAlibaba94.412026-08-23
24Llama 3.3 Instruct Turbo (70B)Meta94.212026-07-02
25Granite 4.1 30BIBM94.212026-06-15
26Llama 3.1 Instruct Turbo (70B)Meta93.812026-07-02
27Llama 3.2 Vision Instruct Turbo (90B)Meta93.612026-07-02
28Kimi K2 BaseMoonshot93.5112026-08-23
29GPT4OpenAI93.232026-08-12
30Mistral Large 2Mistral9322026-08-23
31Hunyuan LargeTencent92.812026-05-31
32DeepSeek-V4-Pro-BaseDeepSeek92.632026-06-05
33Granite 4.1 8BIBM92.512026-06-15
34Nova Micro (Non-Reasoning)Amazon92.332026-08-23
35Claude 3 SonnetAnthropic92.322026-08-23
36MiMo V2 Flash BaseXiaomi92.342026-06-13
37Qwen2 Instruct (72B)Alibaba9222026-08-23
38NVIDIA Nemotron 3 Nano 30B A3BNVIDIA91.412026-08-11
39Llama 3.1 Nemotron Instruct 70BNVIDIA91.412026-08-23
40DeepSeek-V3.1-BaseDeepSeek91.442026-06-13
41Nemotron 3.5 LightningNVIDIA91.312026-08-11
42NVIDIA-Nemotron-3.5-Lightning-30B-A3B-Base-BF16NVIDIA91.312026-08-24
43Mistral Large 3 675B Base 2512Mistral91.212026-06-05
44DeepSeek-V3.2-Exp-BaseDeepSeek91.152026-06-13
45Qwen 2.5 Coder 32B InstructAlibaba91.112026-08-23
46DeepSeek-V4-Flash-BaseDeepSeek90.832026-06-05
47Nemotron 3 Super 120B A12BNVIDIA90.712026-08-11
48Palmyra-X-004Writer90.512026-07-02
49Qwen3.5 35B A3BAlibaba90.112026-08-11
50GLM-4.5-BaseZ.ai90.152026-06-05
51Gemma 3 4BGoogle89.922026-08-23
52Claude 3 HaikuAnthropic88.922026-08-23
53Phi 3.5 MoE InstructMicrosoft88.712026-08-23
54Qwen2.5 Omni 7BAlibaba88.712026-08-23
55Phi 4 Mini InstructMicrosoft88.612026-08-23
56Step3.5 Flash BaseStepFun88.222026-06-05
57DeepSeek-V3-BaseDeepSeek88.222026-06-05
58NVIDIA-Nemotron-3-Ultra-550B-A55B-Base-BF16NVIDIA88.112026-08-24
59Nemotron 3 Ultra 550B A55B BaseNVIDIA88.112026-06-05
60Claude 2Anthropic8832026-08-12
61Phi 3 Medium 4K InstructMicrosoft87.822026-07-03
62Solar ProUpstage87.112026-07-02
63Jamba 1.5 LargeAI218722026-08-23
64Phi 3.5 Mini InstructMicrosoft86.212026-08-23
65Qwen3 1.7BAlibaba85.622026-08-09
66lfm-2.5-1.2b-thinking:freeLiquid AI85.612026-08-09
67Claude 1.3Anthropic85.212026-07-17
68GPT-4o miniOpenAI84.312026-07-02
69MAI-Thinking-1Microsoft8412026-08-14
70Granite 4.1 30B BaseIBM83.822026-06-15
71Llama 3.1 70B InstructMeta83.712026-05-31
72MiMo V2.5 BaseXiaomi83.332026-06-05
73Llama 3.1 8B InstructMeta83.212026-06-11
74Palmyra X V3 (72B)Writer83.112026-07-02
75Qwen2.5 Instruct Turbo (7B)Alibaba8312026-07-02
76Amazon Nova LiteAmazon82.912026-07-02
77LFM2 2.6BLiquid AI82.412026-08-09
78GPT-4 TurboOpenAI82.422026-07-02
79Qwen2 7B InstructAlibaba82.312026-08-23
80Llama 3.2 Vision Instruct Turbo (11B)Meta82.312026-07-02
81Granite 3.2 8B InstructIBM81.712026-06-11
82Claude 3.5 HaikuAnthropic81.522026-07-02
83Gemma 2 Instruct (27B)Google81.212026-07-02
84Granite 3.3 8B InstructIBM80.922026-08-23
85Llama 3 (70B)Meta80.512026-07-02
86Llama 3.1 Instruct Turbo (8B)Meta79.812026-07-02
87DeepSeek-ChatDeepSeek79.512026-07-02
88DeepSeek-V2DeepSeek79.212026-05-31
89Granite 3.1 8B InstructIBM79.212026-06-11
90Qwen2.5 3BAlibaba79.122026-06-05
91Gemini 1.5 Flash (001)Google78.522026-07-02
92DeepSeek-R1-Distill-Qwen-7BDeepSeek78.512026-06-11
93Mistral NemoMistral78.212026-07-02
94Llama 3.2 3B InstructMeta77.742026-08-23
95MoonlightMoonshot77.422026-06-12
96Gemma 4 26B A4BGoogle7712026-08-11
97Arctic InstructSnowflake76.812026-07-02
98Gemma 2 Instruct (9B)Google76.222026-08-23
99Jamba 1.5 MiniAI2175.822026-08-23
100Gemma 2 27B ITGoogle7412026-08-23

Best tracked score per model (default configuration; source-attributed and verification-tiered). Open a model for its full benchmark surface, provenance and pricing.