Skip to content

Benchmark

GSM8K

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

Models and systems measured
104
Scores recorded
104
Sources
5
IQ weight
0.30

Score distribution

74%
Mean
87.6%
Median
28%
Std dev
25.2%
10th percentile
95.1%
90th percentile
0%
Min
0%99.6%
The band covers the middle 80% of scores. The line marks the median.

This benchmark has a weight of 0.30 for chat models, measured over 97 models. The weight is how much it counts toward llminfo IQ. We measure it from the data, not from the raw scores.

Leaderboard

#Model or systemScore
51gemma-2-9b· Epoch AI Benchmarking Hub data84.9%
52Llama 3.1 8B Instruct· OpenEvals Leaderboard data· Reported by the model's developer. The source relaying it did not verify it.
Values from 2 sources
84.5%
53LFM2-8B-A1B· Hugging Face Official Benchmarks data· Reported by the model's developer. The source relaying it did not verify it. claim84.4%
54mistral-nemo-base-2407· Epoch AI Benchmarking Hub data84.2%
55Qwen2.5-Coder-7B· Epoch AI Benchmarking Hub data83.9%
56gemini-1.5-flash-001· Epoch AI Benchmarking Hub data82.4%
57Qwen2 7B Instruct· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim82.3%
58Qwen2.5-Coder-3B-Instruct· Epoch AI Benchmarking Hub data80.7%
59Qwen 2 (7B)· Hugging Face Official Benchmarks data· Reported by the model's developer. The source relaying it did not verify it. claim79.9%
60Text Davinci 003· Epoch AI Benchmarking Hub data78.2%
61Llama 3.2 3B· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim77.7%
62Yi 34B Chat· Epoch AI Benchmarking Hub data76%
63Jamba 1.5 Mini· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim75.8%
64qwen2p5-coder-3b· Epoch AI Benchmarking Hub data75.7%
65Mixtral 8X7b V0.1· Epoch AI Benchmarking Hub data74.4%
66Gemma 2 27B IT· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim74%
67Command R Plus· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim70.7%
68Gemma 2 9B IT· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim68.6%
69Yi-34B· Epoch AI Benchmarking Hub data67.2%
70deepseek-coder-v2-lite-base· Epoch AI Benchmarking Hub data67.1%
71qwen2p5-coder-1p5b· Epoch AI Benchmarking Hub data65.8%
72Gemma 3 1b it· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim62.8%
73Qwen-14B-Chat· Epoch AI Benchmarking Hub data61.2%
74Llama 2 70B Chat· Epoch AI Benchmarking Hub data58.7%
75GPT-3.5 Turbo· Epoch AI Benchmarking Hub data57.8%
76Starcoder2 15b· Epoch AI Benchmarking Hub data57.7%
77Mistral 7B v0.1· Epoch AI Benchmarking Hub data54.4%
78gemma-7b· Epoch AI Benchmarking Hub data46.4%
79Text Davinci 002· Epoch AI Benchmarking Hub data41.5%
80Llama 2 13B Chat· Epoch AI Benchmarking Hub data36.9%
81Mistral 7B Instruct v0.2· Epoch AI Benchmarking Hub data35.4%
82qwen2p5-coder-0p5b· Epoch AI Benchmarking Hub data34.5%
83Llama 2 13b· Epoch AI Benchmarking Hub data34.3%
84starcoder2-7b· Epoch AI Benchmarking Hub data32.7%
85Yi-6B· Epoch AI Benchmarking Hub data32.5%
86ChatGLM2 6B· Epoch AI Benchmarking Hub data32.4%
87ERNIE 4.5· LLM Stats data· Reported by the model's developer. The source relaying it did not verify it. claim25.2%
88starcoder2-3b· Epoch AI Benchmarking Hub data21.6%
89LLaMA 13B· Epoch AI Benchmarking Hub data20.6%
90gemma-2b· Epoch AI Benchmarking Hub data17.7%
91Llama 2 7b· Epoch AI Benchmarking Hub data16.7%
92davinci· Epoch AI Benchmarking Hub data9%
93curie· Epoch AI Benchmarking Hub data1.6%
94babbage· Epoch AI Benchmarking Hub data0.7%
95ada· Epoch AI Benchmarking Hub data0.6%
96Text Curie 001· Epoch AI Benchmarking Hub data0.6%
97Text Ada 001· Epoch AI Benchmarking Hub data0.4%
98Text Babbage 001· Epoch AI Benchmarking Hub data0%

Showing 51–98 of 98 results.

Not ranked

The source marks these results as descriptive, not capability scores, so we list them by name instead of ranking them.

Model or systemScore
DeepSeek V4.1 Flash BaseEM · 8-shot· Reported by DeepSeek· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it.93%
MiMo V2 Flash Baseofficial Base table · 8-shot· Reported by Xiaomi MiMo· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it.92.3%
Qwen3.5-397B-A17B-BaseQwen3.5 release/model card · base-model table· Reported by Qwen· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it.93.7%
Qwen3.7-Plus-BaseQwen3.8-Flash-Next release/model card · base-model table· Reported by Qwen· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it.93%
Qwen3.8-27B-BaseQwen3.8-Flash-Next release/model card · base-model table· Reported by Qwen· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it.93.2%
Qwen3.8-Flash-Next-BaseQwen3.8-Flash-Next release/model card · base-model table· Reported by Qwen· via BenchBoard docs· Reported by the model's developer. The source relaying it did not verify it.93.3%

Data credits: Epoch AI Benchmarking Hub, CC BY 4.0 (licence) · Data by LLM Stats · Hugging Face official benchmark leaderboards; licence varies by benchmark · OpenEvals/leaderboard-data, MIT (licence) · 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.