MATH Level 5

Reasoning

Source ↗ Higher is better

Rank	Model	Provider	Score
1	nvidia/AceMath-72B-Instruct	NVIDIA	71.5%
2	nvidia/AceMath-7B-Instruct	NVIDIA	63.4%
3	nvidia/AceInstruct-72B	NVIDIA	62.6%
4	Qwen/Qwen2.5-32B-Instruct	Alibaba	62.5%
5	Qwen/Qwen2.5-Math-72B-Instruct	Alibaba	62.4%
6	maldv/Awqward2.5-32B-Instruct	maldv	62.3%
7	Saxo/Linkbricks-Horizon-AI-Avengers-V6-32B	Saxo	62.2%
8	Saxo/Linkbricks-Horizon-AI-Avengers-V3-32B	Saxo	61.8%
9	tanliboy/lambda-qwen2.5-32b-dpo-test	tanliboy	61.0%
10	Saxo/Linkbricks-Horizon-AI-Avengers-V1-32B	Saxo	60.3%
11	huihui-ai/Qwen2.5-72B-Instruct-abliterated	huihui-ai	60.1%
12	Qwen/Qwen2.5-72B-Instruct	Alibaba	59.8%
13	zetasepic/Qwen2.5-32B-Instruct-abliterated-v2	zetasepic	59.5%
14	CombinHorizon/huihui-ai-abliterated-Qwen2.5-32B-Inst-BaseMerge-TIES	CombinHorizon	59.4%
15	MaziyarPanahi/calme-2.1-qwen2.5-72b	MaziyarPanahi	59.1%
16	MaziyarPanahi/calme-2.2-qwen2.5-72b	MaziyarPanahi	58.9%
17	CombinHorizon/zetasepic-abliteratedV2-Qwen2.5-32B-Inst-BaseMerge-TIES	CombinHorizon	58.5%
18	Qwen/Qwen2.5-Math-7B-Instruct	Alibaba	58.1%
19	prithivMLmods/Gauss-Opus-14B-R999	prithivMLmods	57.6%
20	deepseek-ai/DeepSeek-R1-Distill-Qwen-14B	DeepSeek	57.0%
21	prithivMLmods/Sombrero-Opus-14B-Sm1	prithivMLmods	56.6%
22	Saxo/Linkbricks-Horizon-AI-Avengers-V2-32B	Saxo	56.6%
23	prithivMLmods/Viper-Coder-v1.6-r999	prithivMLmods	56.6%
24	jpacifico/Chocolatine-14B-Instruct-DPO-v1.3	jpacifico	56.2%
25	rootxhacker/Apollo-70B	rootxhacker	56.1%
26	TIGER-Lab/Qwen2.5-Math-7B-CFT	TIGER-Lab	55.7%
27	Lunzima/NQLSG-Qwen2.5-14B-MegaFusion-v8	Lunzima	55.6%
28	Triangle104/DS-R1-Distill-Q2.5-14B-Harmony_V0.1	Triangle104	55.5%
29	Qwen/Qwen2-Math-72B-Instruct	Alibaba	55.4%
30	braindao/Qwen2.5-14B-Instruct	braindao	55.3%
31	prithivMLmods/Coma-II-14B	prithivMLmods	55.1%
32	CombinHorizon/huihui-ai-abliteratedV2-Qwen2.5-14B-Inst-BaseMerge-TIES	CombinHorizon	54.8%
33	Qwen/Qwen2.5-14B-Instruct	Alibaba	54.8%
34	mlabonne/BigQwen2.5-52B-Instruct	mlabonne	54.8%
35	Saxo/Linkbricks-Horizon-AI-Avengers-V5-32B	Saxo	54.6%
36	prithivMLmods/Viper-Coder-v1.1	prithivMLmods	54.6%
37	tanliboy/lambda-qwen2.5-14b-dpo-test	tanliboy	54.6%
38	1024m/QWEN-14B-B100	1024m	54.4%
39	Goekdeniz-Guelmez/Josiefied-Qwen2.5-14B-Instruct-abliterated-v4	Goekdeniz-Guelmez	54.2%
40	rombodawg/Rombos-LLM-V2.5-Qwen-72b	rombodawg	54.2%
41	tomasmcm/sky-t1-coder-32b-flash	tomasmcm	54.2%
42	nbeerbower/Qwen2.5-Gutenberg-Doppel-14B	nbeerbower	54.2%
43	Lunzima/NQLSG-Qwen2.5-14B-MegaFusion-v8.7	Lunzima	54.1%
44	fluently-lm/FluentlyLM-Prinum	fluently-lm	54.0%
45	suayptalha/Lamarckvergence-14B	suayptalha	54.0%
46	YOYO-AI/ZYH-LLM-Qwen2.5-14B-V4	YOYO-AI	53.9%
47	FINGU-AI/RomboUltima-32B	FINGU-AI	53.9%
48	YOYO-AI/Qwen2.5-14B-it-restore	YOYO-AI	53.7%
49	Lunzima/NQLSG-Qwen2.5-14B-MegaFusion-v8.9	Lunzima	53.7%
50	v000000/Qwen2.5-Lumen-14B	v000000	53.6%

Access this data

JSON GET /api/v1

[
  {
    "rank": 1,
    "model": "nvidia/AceMath-72B-Instruct",
    "provider": "NVIDIA",
    "score": "71.45015105740181"
  },
  {
    "rank": 2,
    "model": "nvidia/AceMath-7B-Instruct",
    "provider": "NVIDIA",
    "score": "63.368580060422964"
  },
  {
    "rank": 3,
    "model": "nvidia/AceInstruct-72B",
    "provider": "NVIDIA",
    "score": "62.61329305135952"
  }
]

Curl

$ curl -s https://rightrank.ai/api/v1/rankings/math-level-5.json | jq '.[:3]'

CLI brew install rightrank

$ rightrank top math-level-5 --limit 5
$ rightrank compare claude-opus-4.6 gpt-5.4-pro --category math-level-5
$ rightrank value math-level-5  # best performance per dollar

Claude Code Skill Skill

# In Claude Code, use the RightRank skill:
/rightrank best model for math-level-5
/rightrank compare top 5 by value
/rightrank cheapest model scoring above 80 on math-level-5