Swallow - Post-trained English - MATH-500 Leaderboard

Swallow - Post-trained English - MATH-500: leaderboard

Metric: Accuracy (%). Source: swallow-llm.github.io. 68 models tracked.

#	Model	Score
1	GPT-5.4 (Medium)	99.2
2	GPT-5 (Medium)	99
3	GPT-OSS-Swallow-120B-RL-v0.1 (Thinking)	99
4	Gemma 4 31B (IT) (Thinking)	99
5	Qwen 3 235B A22B 2507 (Thinking)	98.6
6	gemma-4-26B-A4B-it (Thinking)	98.6
7	GPT-OSS-Swallow-20B-RL-v0.1 (Thinking)	98.4
8	Qwen 3.5 35B A3B (Thinking)	98.4
9	Qwen 3 235B A22B 2507 Instruct	98.2
10	Qwen 3 Next 80B A3B (Thinking)	98
11	Qwen 3 Next 80B A3B Instruct	98
12	O3 (2025-04-16) (Medium)	97.8
13	Qwen 3 32B (Thinking)	97.6
14	Qwen 3 8B (Thinking)	97.6
15	Qwen 3.5 122B A10B (Thinking)	97.4