Skip to content

Benchmark

IMOAnswerBench

Math benchmarks measure how well models solve maths problems. Scores are percentages. Higher is better.

Models measured
25
Scores recorded
44
Sources
2
IQ weight
2.35

Score distribution

82.6%
Mean
83.9%
Median
9.3%
Std dev
78.3%
10th percentile
90%
90th percentile
41.9%
Min
41.9%92.3%
The band covers the middle 80% of scores. The line marks the median.

This benchmark has a weight of 2.35 for chat models, measured over 25 models. The weight is how much it counts toward llminfo IQ. We measure it from the data, not from the raw scores.

Leaderboard

#ModelScore
1Nemotron 3 Ultra· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim92.3%
2Nemotron 3 UltraNeMo Evaluator SDK · with tools· Reported by NVIDIA· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it.92.3%
3GLM 5.2· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim91%
4GLM 5.2GPT-5.5 medium judge· Reported by Z.ai· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it.91%
5Hy3 (Deep Infra)· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim90%
6Hy3 (Deep Infra)Hy3 official Benchmark Appendix · highest reasoning tier unless noted· Reported by Tencent Hunyuan· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it.90%
7Qwen3.7 Max· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim90%
8Qwen3.7 MaxQwen3.7-Max release · recommended reasoning-effort=xhigh system prompt· Reported by Qwen· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it.90%
9Step 3.5 FlashStep 3.5 technical report · Table 5 · PaCoRe· Reported by StepFun· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it.88.8%
10Nemotron 3 UltraNeMo Evaluator SDK · no tools· Reported by NVIDIA· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it.88.6%
11DeepSeek V4 FlashDeepSeek V4 official model card · Think Max· Reported by DeepSeek· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it.88.4%
12Step 3.5 FlashStep 3.5 technical report · Tables 12–13 · Python tool· Reported by StepFun· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it.86.7%
13Kimi K2.6· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim86%
14Kimi K2.6Kimi K2.6 official benchmark table · thinking · temp 1.0 · top_p 1.0 · 262,144 context· Reported by Moonshot AI· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it.86%
15Qwen3.7 Plus· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim86%
16Qwen3.7 PlusQwen3.7-Plus release · recommended reasoning-effort=xhigh system prompt· Reported by Qwen· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it.86%
17Step 3.5 Flash· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim85.4%
18Step 3.5 FlashStep 3.5 Flash official headline table · vanilla· Reported by StepFun· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it.85.4%
19DeepSeek V4 Flash· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim85.1%
20DeepSeek V4 FlashDeepSeek V4 official model card · Think High· Reported by DeepSeek· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it.85.1%
21DeepSeek V3.2 SpecialeSpeciale high-compute reasoning · temperature 1.0 · 128K context · pass@1· Reported by DeepSeek· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it.84.5%
22Qwen3 Max· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim83.9%
23GLM 5.1· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim83.8%
24GLM 5.1GLM-5.1 official release table· Reported by Z.ai· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it.83.8%
25Qwen3.6 Plus· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim83.8%
26Qwen3.6 PlusQwen official release table · thinking mode unless benchmark requires otherwise· Reported by Alibaba Qwen· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it.83.8%
27Ling 3.0 Flash· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim83.7%
28GLM 5GLM-5 official release table· Reported by Z.ai· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it.82.5%
29GLM 4.7· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim82%
30Kimi K2.5setting not named by the source· Reported by Moonshot AI· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it.81.8%
31Kimi K2.5· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim81.8%
32Qwen3.5 397B A17B· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim80.9%
33Qwen3.5 397B A17BQwen3.5 release/model card · published setting· Reported by Qwen· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it.80.9%
34Qwen3.6 27B· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim80.8%
35Qwen3.6 27BQwen official release table · thinking mode unless benchmark requires otherwise· Reported by Alibaba Qwen· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it.80.8%
36Qwen3.6 35B A3B· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim78.9%
37Qwen3.6 FlashQwen3.6-35B-A3B official model card · hosted API name Qwen3.6-Flash· Reported by Alibaba Qwen· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it.78.9%
38Kimi K2 Thinking· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim78.6%
39LongCat-Flash-Thinking· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim78.6%
40DeepSeek V3.2· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim78.3%
41DeepSeek V3.2Thinking mode · temperature 1.0 · 128K context · custom step-by-step / boxed-answer prompt· Reported by DeepSeek· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it.78.3%
42Qwen3.5 Omni PlusQwen3.5-Omni release · no-thinking setting· Reported by Qwen, Qwen Team· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it.65.5%
43Qwen3.5 Omni FlashQwen3.5-Omni release · no-thinking setting· Reported by Qwen, Qwen Team· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it.51.5%
44DeepSeek V4 FlashDeepSeek V4 official model card · Non-Think· Reported by DeepSeek· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it.41.9%

Data credits: BenchBoard by SYuan03 (github.com/SYuan03/benchboard). MIT covers its code and data organization; no results licence is inferred. Numeric relays retain the original stating-source links. · Data by LLM Stats