AIME leaderboard

Problems from the American Invitational Mathematics Examination, a qualifier for the US Math Olympiad. Gemini 3.1 Pro Preview leads with 98.13%.

Models
80
Leader
Gemini 3.1 Pro Preview
Source
Vals AI
As of
Oct 24, 2026
  1. 1Gemini 3.1 Pro Preview98.1%
  2. 2GPT-5.296.9%
  3. 3Muse Spark96.9%
  4. 4Gemini 3 Pro96.7%
  5. 5GPT-5.496.7%
  6. 6Grok 4.2096.5%
  7. 7Claude Opus 4.796.3%
  8. 8Gemini 3 Flash95.6%
  9. 9Claude Opus 4.695.6%
  10. 10GPT-5.4 mini95.6%
AIME, accuracy against cost per test
Open weightsProprietaryBest for the price
0%25%50%75%100%$.1$1$10Cost per test, log scale →Gemini 3.1 Pro PreviewGemini 3 Flashgpt-oss-120bgpt-oss-20b

Full leaderboard

RankModelLabAccuracyCost / testHarnessPrice in / out
1Gemini 3.1 Pro PreviewGoogle98.13%$2 / $12—$2 / $12
2GPT-5.2OpenAI96.88%$1.75 / $14—$1.75 / $14
3Muse SparkMeta96.88%N/A——
4Gemini 3 ProGoogle96.68%$2 / $12——
5GPT-5.4OpenAI96.67%$2.5 / $15—$2.50 / $15
6Grok 4.20xAI96.46%$2 / $6—$1.25 / $2.50
7Claude Opus 4.7Anthropic96.25%$5 / $25—$5 / $25
8Gemini 3 FlashGoogle95.63%$0.5 / $3—$0.50 / $3
9Claude Opus 4.6Anthropic95.63%$5 / $25—$5 / $25
10GPT-5.4 miniOpenAI95.63%$0.75 / $4.5—$0.75 / $4.50
11Kimi K2.5Moonshot AI95.63%$0.6 / $3—$0.50 / $2.50
13Qwen3.6-PlusAlibaba94.58%$0.5 / $3—$0.33 / $1.95
14GPT-5OpenAI93.37%$1.25 / $10—$1.25 / $10
15GPT-5.1OpenAI93.33%$1.25 / $10—$1.25 / $10
16GLM-4.7Z.ai93.33%$0.6 / $2.2—$0.54 / $1.98
17GLM-4.6Z.ai92.71%$0.6 / $2.2—$0.55 / $2.20
18gpt-oss-120bOpenAI92.60%$0.15 / $0.6—$0.15 / $0.60
19Qwen3.5-FlashAlibaba92.50%$0.1 / $0.4—$0.065 / $0.26
20Claude Sonnet 4.6Anthropic92.29%$3 / $15—$3 / $15
21Grok 4.1 FastxAI91.88%$0.2 / $0.5——
22GLM-5.1Z.ai91.88%$1 / $3.2—$1.40 / $4.40
23GLM-5Z.ai91.67%$1 / $3.2—$1 / $3.20
24GPT-5 miniOpenAI91.46%$0.25 / $2—$0.25 / $2
26MiniMax-M2.7MiniMax91.04%$0.3 / $1.2—$0.30 / $1.20
27Grok 4xAI90.56%$3 / $15——
28GPT-5.4 nanoOpenAI88.75%$0.2 / $1.25—$0.20 / $1.25
29MiniMax-M2.5MiniMax88.75%$0.3 / $1.2—$0.30 / $1.20
30Claude Sonnet 4.5Anthropic88.19%$3 / $15—$3 / $15
31GLM-4.5Z.ai86.67%$0.6 / $2.2—$0.60 / $2.20
32o3-miniOpenAI86.46%$1.1 / $4.4—$1.10 / $4.40
33Qwen 3.5 PlusAlibaba86.04%$0.4 / $2.4—$0.30 / $1.80
34gpt-oss-20bOpenAI86.04%$0.07 / $0.3—$0.018 / $0.09
35Gemini 2.5 ProGoogle85.83%$1.25 / $10—$1.25 / $10
36Kimi K2 ThinkingMoonshot AI85.42%$0.6 / $2.5—$0.60 / $2.50
37o3OpenAI85.28%$2 / $8—$2 / $8
40Qwen3-235B-A22BAlibaba83.96%$0.22 / $0.88—$0.09 / $0.55
41o4-miniOpenAI83.67%$1.1 / $4.4—$1.10 / $4.40
42Magistral MediumMistral AI83.54%$2 / $5——
43Gemini 3.1 Flash-Lite PreviewGoogle83.33%$0.25 / $1.5—$0.25 / $1.50
44Claude Haiku 4.5Anthropic82.71%$1 / $5—$1 / $5
45GPT-5 nanoOpenAI81.18%$0.05 / $0.4—$0.05 / $0.40
49MiniMax-M2.1 PreviewMiniMax77.92%$0.3 / $1.2—$0.30 / $1.20
50Claude Opus 4.5Anthropic76.88%$5 / $25—$5 / $25
52DeepSeek-R1DeepSeek73.96%$1.35 / $5.4—$0.50 / $2.15
53o1OpenAI71.46%$15 / $60—$15 / $60
54Grok 3 Mini ReasoningxAI70.63%$0.3 / $0.5——
55DeepSeek-V3.2-ExpDeepSeek64.79%$0.56 / $1.68—$0.28 / $0.42
56Kimi K2 InstructMoonshot AI62.71%$1 / $3——
57Qwen3-Max-PreviewAlibaba60.69%$1.2 / $6——
58Grok 3 PreviewxAI58.75%$3 / $15——
63GPT-4.1 miniOpenAI49.38%$0.4 / $1.6—$0.40 / $1.60
65Claude Opus 4.1Anthropic44.24%$15 / $75—$15 / $75
67Mistral MediumMistral AI42.29%$0.4 / $2——
69Claude Opus 4Anthropic41.25%$15 / $75——
70GPT-4.1OpenAI39.58%$2 / $8—$2 / $8
71Claude Sonnet 4Anthropic38.54%$3 / $15——
72Grok 4 FastxAI33.33%$0.2 / $0.5——
73Gemini 2.0 FlashGoogle29.79%$0.1 / $0.4——
74DeepSeek-V3DeepSeek27.50%$0.9 / $0.9—$0.29 / $1.14
75Grok 4.1 Fast Non-ReasoningxAI27.08%$0.2 / $0.5——
76GPT-4.1 nanoOpenAI26.46%$0.1 / $0.4—$0.10 / $0.40
77Gemini 2.5 Flash PreviewGoogle26.25%$0.1 / $0.4——
78Llama 4 MaverickFireworks AI25.21%$0.22 / $0.88—$0.19 / $0.65
79Claude 3.7 SonnetAnthropic22.29%$3 / $15——
80Llama 4 ScoutMeta18.96%$0.18 / $0.59—$0.10 / $0.30
81Gemini 1.5 ProGoogle18.75%$1.25 / $5——
82Gemini 1.5 FlashGoogle17.29%$0.075 / $0.3——
83Llama 3.3 Instruct TurboMeta16.04%$0.88 / $0.88——
84Grok 2xAI15.21%$2 / $10——
86Command ACohere13.33%$2.5 / $10—$0.30 / $1.50
87GPT-4oOpenAI11.88%$2.5 / $10—$2.50 / $10
88GPT-4o miniOpenAI11.46%$0.15 / $0.6—$0.15 / $0.60
89Claude 3.5 SonnetAnthropic10.00%$3 / $15——
90Llama 3.3 Nemotron SuperMeta9.38%N/A——
91Mistral LargeMistral AI9.17%$2 / $6—$0.50 / $1.50
92Mistral SmallMistral AI5.63%$0.2 / $0.6——
93Magistral Small 1.2Mistral AI3.54%$0.075 / $0.3——
94Claude 3.5 HaikuAnthropic3.33%$0.8 / $4——
95Jamba 1.6 MiniAI21 Labs0.42%$0.2 / $0.4——
96Jamba 1.6 LargeAI21 Labs0.42%$2 / $8——

Related leaderboards

Sources

  • Vals AI, as of Oct 24, 2026
  • Provider list prices, as of Oct 24, 2026

Every number on this page comes from these sources; models without a published number show none. Download the data. Cite as “Nexus Explorer, nexus.inc/models”.