Back to home
Reasoning
How well can each AI model solve complex logical and mathematical problems? Ranked by Artificial Analysis Math Index.
RankModelScore
1
23
4
567891011
121314151617
18
19
2021222324252627282930
GPT-5openai/gpt-5-2
99.0
GPT-5openai/gpt-5-codex
98.7
Gemini 3 Flashgoogle/gemini-3-flash-reasoning
97.0
DeepSeek V3deepseek/deepseek-v3-2-speciale
96.7
X
MiMo V2 Flashxiaomi/mimo-v2-flash-reasoning
96.3
GPT-5openai/gpt-5-1-codex
95.7
Gemini 3 Progoogle/gemini-3-pro
95.7
Z
GLM 4.7zai/glm-4-7
95.0
K
KAT Coder Pro V1kwaikat/kat-coder-pro-v1
94.7
K
Kimi K2kimi/kimi-k2-thinking
94.7
Nova 2.0 Liteaws/nova-2-0-lite-reasoning
94.3
GPT-5openai/gpt-5
94.3
GPT-5openai/gpt-5-1
94.0
gpt oss 120bopenai/gpt-oss-120b
93.4
S
Grok 4xai/grok-4
92.7
DeepSeek V3deepseek/deepseek-v3-2-reasoning
92.0
Claude Opus 4.5anthropic/claude-opus-4-5-thinking
91.3
NVIDIA Nemotron 3 Nano 30B A3Bnvidia/nvidia-nemotron-3-nano-30b-a3b-reasoning
91.0
Qwen3 235B A22B 2507alibaba/qwen3-235b-a22b-instruct-2507-reasoning
91.0
o4 Miniopenai/o4-mini
90.7
L
K EXAONElg/k-exaone
90.3
Nova 2.0 Omniaws/nova-2-0-omni-reasoning-medium
89.7
DeepSeek V3deepseek/deepseek-v3-1-terminus-reasoning
89.7
DeepSeek V3deepseek/deepseek-v3-1-reasoning
89.7
S
Grok 4xai/grok-4-fast-reasoning
89.7
gpt oss 20bopenai/gpt-oss-20b
89.3
S
Grok 4xai/grok-4-1-fast-reasoning
89.3
I
Ring 1Tinclusionai/ring-1t
89.3
Nova 2.0 Proaws/nova-2-0-pro-reasoning-medium
89.0
o3openai/o3
88.3
194 models tested