llmrun
Reasoning

SimpleBench Leaderboard

SimpleBench is a set of everyday, common-sense and trick questions that humans answer easily but language models often get wrong. It probes basic reasoning and robustness rather than specialist knowledge.

Source: epoch19 open models ranked+71 proprietaryData through Jul 2026

Open modelsAll models

Open models ranked on SimpleBench

# shows rank among open models / rank overall (including proprietary).

#ModelScore
1 / 22DeepSeek V4 Pro · 861.6B
61.2%
2 / 30Kimi K2.7 Code · 1058.6B
57.9%
3 / 35GLM 5.1 · 753.9B
55.1%
4 / 38GLM 5 · 753.9B
53.2%
5 / 41DeepSeek V3.2 Speciale · 685.4B
52.6%
6 / 45GLM 4.7 · 358.3B
47.7%
7 / 47Kimi K2.5 · 1058.6B
46.8%
8 / 52MiniMax M3 · 427.0B
45.8%
9 / 59DeepSeek R1 0528 · 684.5B
40.8%
10 / 61DeepSeek V3.1 · 684.5B
40.0%
11 / 68Qwen3 235B A22B · 235.1B
31.0%
12 / 69DeepSeek R1 · 684.5B
30.9%
13 / 71Llama 4 Maverick 17B 128E Instruct · 401.6B
27.7%
14 / 73DeepSeek v3 0324 · 684.5B
27.2%
15 / 76Kimi K2 Instruct · 1026.5B
26.3%
16 / 79Llama 3.1 405B Instruct · 405.9B
23.0%
17 / 83GPT OSS 120B · 120.4B
22.1%
18 / 84Llama 3.3 70B Instruct · 70.6B
19.9%
19 / 85DeepSeek v3 · 684.5B
18.9%

Score vs model size

Which models give the most quality for their size — the ones worth running locally.

100B1Tmodel size (log scale) →61.2%18.9%Kimi K2.7 Code · 1.1T · 57.9%GLM 5 · 754B · 53.2%Kimi K2.5 · 1.1T · 46.8%MiniMax M3 · 427B · 45.8%DeepSeek R1 0528 · 685B · 40.8%DeepSeek V3.1 · 685B · 40.0%DeepSeek R1 · 685B · 30.9%Llama 4 Maverick 17B 128E Instruct · 402B · 27.7%DeepSeek v3 0324 · 685B · 27.2%Kimi K2 Instruct · 1T · 26.3%Llama 3.1 405B Instruct · 406B · 23.0%DeepSeek v3 · 685B · 18.9%Llama 3.3 70B Instruct · 71B · 19.9%Llama 3.3 70B InstructGPT OSS 120B · 120B · 22.1%GPT OSS 120BQwen3 235B A22B · 235B · 31.0%Qwen3 235B A22BGLM 4.7 · 358B · 47.7%GLM 4.7DeepSeek V3.2 Speciale · 685B · 52.6%DeepSeek V3.2 SpecialeGLM 5.1 · 754B · 55.1%GLM 5.1DeepSeek V4 Pro · 862B · 61.2%DeepSeek V4 Pro
Each dot is a model. Up = higher score, left = smaller (easier to run locally). The dashed line marks the efficiency frontier — the best score you can get at each size or smaller.
  • Llama 3.3 70B Instruct, 71B, score 19.9% — on the efficiency frontier (best score at its size or smaller).
  • GPT OSS 120B, 120B, score 22.1% — on the efficiency frontier (best score at its size or smaller).
  • Qwen3 235B A22B, 235B, score 31.0% — on the efficiency frontier (best score at its size or smaller).
  • GLM 4.7, 358B, score 47.7% — on the efficiency frontier (best score at its size or smaller).
  • DeepSeek V3.2 Speciale, 685B, score 52.6% — on the efficiency frontier (best score at its size or smaller).
  • GLM 5.1, 754B, score 55.1% — on the efficiency frontier (best score at its size or smaller).
  • DeepSeek V4 Pro, 862B, score 61.2% — on the efficiency frontier (best score at its size or smaller).

SimpleBench: frequently asked questions

What is the best open LLM on SimpleBench?
DeepSeek V4 Pro is the top open model on SimpleBench, scoring 61.2%. Among all models tested — including proprietary ones — it ranks #22. The top model overall is Claude Fable 5 Max (Anthropic) at 81.9%.
Can open models match proprietary models on SimpleBench?
Not quite on SimpleBench: the strongest proprietary model (Claude Fable 5 Max) scores 81.9%, ahead of the best open model (DeepSeek V4 Pro) at 61.2% — but you can run the open one yourself.

Scores aggregated from epoch. llmrun does not run this benchmark — see the source for methodology, or the about benchmarks for what it measures.