Atlas

Benchmarks

← All benchmarks

Kangaroo 2025 11-12

Math · 2025-10-20

This MathArena track evaluates models on the Kangaroo 2025 11-12 problem set. Scores report the percentage of problems answered correctly.

Top models (higher is better)

ModelScore
Opus 4.8100.0
GPT-5.498.3
GPT-5.598.3
Kimi K398.3
Gemini 3.1 Pro Preview97.5
Gemini 3 Pro Preview97.5
Gemini 3.6 Flash96.7
Gemini 3 Flash Preview95.8
GPT-5.295.8
GPT-595.0
Gemini 3.5 Flash93.3
GPT-5.191.7
Qwen3.5 397B A17B91.7
Kimi K2.590.0
Gemini 2.5 Pro89.2
Opus 4.687.5
Sonnet 4.587.5
Qwen3-VL-235B-A22B-Instruct85.8
GPT-5 Mini85.0
Grok 4.1 Fast81.7
Grok 4 Fast80.8
GLM-4.5V78.3
Grok 476.7
Loading Atlas data…