MATH-500

35 models

Top 5 Models Performance

Meituan LongCat: longcat-flash-thinking	########################################	99.2
Prism ML: ternary-bonsai-27b	########################################	99.2
Prism ML: ternary-bonsai-2-27b	########################################	98.8
orcarouter: orcacyber-nano-1.0	########################################	98.6
Sarvam AI: sarvam-105b	########################################	98.6

	Rank	Model	Score
🇨🇳	1	Meituan LongCat: longcat-flash-thinking	99.2
❓	2	Prism ML: ternary-bonsai-27b	99.2
❓	3	Prism ML: ternary-bonsai-2-27b	98.8
❓	4	orcarouter: orcacyber-nano-1.0	98.6
🇮🇳	5	Sarvam AI: sarvam-105b	98.6
❓	6	webai-official: twil-8b	98.2
🇨🇳	7	Z.ai: glm-4.5北京智谱华章科技有限公司	98.2
🇨🇳	8	Z.ai: glm-4.5-air北京智谱华章科技有限公司	98.1
🇺🇸	9	NVIDIA Corporation: nvidia-nemotron-nano-9b-v2	97.8
🇦🇪	10	Technology Innovation Institute: falcon-h1r-7b	97.4
🇨🇳	11	Qwen: Qwen3 8B通义千问	96.3
❓	12	diffuspace: clover-35b-a3b-preview	96.2
🇺🇸	13	LiquidAI: lfm-2.5-24b-a2b	96.04
🇺🇸	14	NVIDIA Corporation: NVIDIA Nemotron 3 Nano 4B	95.4
🇨🇳	15	Qwen: Qwen3 4B通义千问	95
🇨🇳	16	OpenBMB: minicpm5-2b	94.6
🇺🇸	17	AMD: sand-math-qwen2.5-32b	93.85
🇺🇸	18	AMD: sand-mathscience-deepseek-qwen32b	93.85
🇨🇳	19	Tencent: youtu-llm-2b腾讯控股有限公司	93.7
❓	20	Ornith: ornith-1.5-9b	92
🇺🇸	21	Hugging Face Smol Models Research: smollm3-3b	91.8
🇨🇳	22	OpenBMB: minicpm5-1b	91.6
🇷🇺	23	Yandex: aliceai-foundation-80b-a3b-baseЯндекс	91.1
🇨🇳	24	Qwen: qwen3-1.7b通义千问	89.8
🇺🇸	25	LiquidAI: lfm-2.5-8b-a1b	88.76
🇷🇺	26	Sber: gigachat3.5-432b-a28bСбер	86
🇨🇳	27	DeepSeek: deepseek-r1-distill-qwen-1.5b深度求索	84.8
🇦🇪	28	Technology Innovation Institute: falcon-h1-1.5b-deep-instruct	77.8
🇦🇪	29	Technology Innovation Institute: falcon-h1-1.5b-instruct	74
🇺🇸	30	Microsoft: phi-4-mini-instruct	69.6
🇺🇸	31	Hugging Face Smol Models Research: qwen3-1.7b-gsm8k	55.2
🇺🇸	32	Google: Gemma 3 4B Instruct	36.4
🇺🇸	33	Arcee AI: trinity-nano-base	36
🇨🇳	34	OpenBMB: bitcpm4-1b	34
❓	35	viesar: gemma-3-4b-opus-reasoning-distill	29.6