MMLU subject subset: Case-style problems in financial accounting, auditing, cost accounting, tax and business law.
unassessed
| Category | knowledge |
|---|---|
| Subcategory | other |
| Page status | active |
| Metric | accuracy |
| Direction | higher_is_better |
| Unit | % |
| Dataset size | 282 |
| Dataset licence | MIT |
| Publisher | UC Berkeley (original); Center for AI Safety (current host) |
Case-style problems in financial accounting, auditing, cost accounting, tax and business law, mixing numeric calculation with rule application. Questions are four-option multiple-choice, drawn from the MMLU test set's "other" subcategory within the benchmark's "other" top-level group, and are graded on the single correct labelled option.
Four-option multiple-choice questions, graded on the single correct labelled option; commonly evaluated 5-shot, consistent with the rest of MMLU.
| Model | Provider | Score | Card as of |
|---|---|---|---|
| Claude Opus 4 | Anthropic | 72.5 | 2026-04 |
| Claude Opus 4.6 | Anthropic | 72.5 | 2026-04 |
| GPT-4.1 | OpenAI | 71.8 | 2026-04 |
| Gemini 2.5 Pro | Google DeepMind | 71.2 | 2026-04 |
| GPT-4o | OpenAI | 70.5 | 2026-04 |
| GPT-4o (2024-05-13) | OpenAI | 70.5 | 2026-04 |
| GPT-4o (2024-08-06) | OpenAI | 70.5 | 2026-04 |
| GPT-4o (2024-11-20) | OpenAI | 70.5 | 2026-04 |
| GPT-4o mini | OpenAI | 70.5 | 2026-04 |
| Claude Sonnet 4 | Anthropic | 69.8 | 2026-04 |
| Claude Sonnet 4.5 | Anthropic | 69.8 | 2026-04 |
| Claude Sonnet 4.5 (latest) | Anthropic | 69.8 | 2026-04 |
| DeepSeek R1 | DeepSeek | 68.2 | 2026-04 |
| DeepSeek R1 0528 | DeepSeek | 68.2 | 2026-04 |
| DeepSeek R1 0528 NVFP4 v2 | NVIDIA | 68.2 | 2026-04 |
| DeepSeek R1 0528 Qwen3 8B | DeepSeek | 68.2 | 2026-04 |
| DeepSeek R1 Distill Llama 70B | DeepSeek | 68.2 | 2026-04 |
| DeepSeek R1 Distill Llama 8B | DeepSeek | 68.2 | 2026-04 |
| DeepSeek R1 Distill Qwen 1.5B | DeepSeek | 68.2 | 2026-04 |
| DeepSeek R1 Distill Qwen 14B | DeepSeek | 68.2 | 2026-04 |
| DeepSeek R1 Distill Qwen 32B | DeepSeek | 68.2 | 2026-04 |
| DeepSeek R1 Distill Qwen 7B | DeepSeek | 68.2 | 2026-04 |
| DeepSeek Reasoner | DeepSeek | 68.2 | 2026-04 |
| Qwen 3 235B Instruct | Cerebras | 67.5 | 2026-04 |
| Qwen3 235B-A22B | Alibaba / Qwen Team | 67.5 | 2026-04 |
| Mixtral 8x22B Instruct v0.1 | Mistral AI | 66.3 | 2026-04 |
| Mistral Large (latest) | Mistral AI | 65.8 | 2026-04 |
| Mistral Large 2.1 | Mistral AI | 65.8 | 2026-04 |
| Mistral Large 3 | Mistral AI | 65.8 | 2026-04 |
| Gemma 4 31B | Google DeepMind | 65.2 | 2026-04 |
| gemma 4 31B it | Google DeepMind | 65.2 | 2026-04 |
| gemma 4 31B it GGUF | Unsloth | 65.2 | 2026-04 |
| Gemma 4 31B IT NVFP4 | NVIDIA | 65.2 | 2026-04 |
| Nous Hermes 2 Yi 34B | Nous Research | 64.9 | 2024-07 |
| Gemma 4 26B | Google DeepMind | 64.5 | 2026-04 |
| Yi 1.5 34B 32K | 01.AI | 64.2 | 2024-07 |
| Llama 3.3 70B Instruct NVFP4 | NVIDIA | 63.8 | 2026-04 |
| Llama-3.3-70B-Instruct | Meta | 63.8 | 2026-04 |
| Meta Llama 3 70B | Meta | 63.8 | 2024-07 |
| Meta Llama 3 70B Instruct | Meta | 63.8 | 2026-04 |
| Meta Llama 3 70B Instruct | Nous Research | 63.8 | 2026-04 |
| Yi 1.5 34B Chat | 01.AI | 63.1 | 2026-04 |
| Yi 1.5 34B Chat 16K | 01.AI | 63.1 | 2024-07 |
| Llama 3.1 70B | Meta | 62.5 | 2026-04 |
| Llama 3.1 70B Instruct | Meta | 62.5 | 2026-04 |
| Yi 1.5 34B | 01.AI | 62.1 | 2026-04 |
| Yi 34B Chat | 01.AI | 61.7 | 2024-07 |
| Yi 34B 200K | 01.AI | 61.3 | 2024-07 |
| phi 4 | Microsoft | 61.2 | 2026-04 |
| Phi 4 mini instruct | Microsoft | 61.2 | 2026-04 |
| Phi 4 multimodal instruct | Microsoft | 61.2 | 2026-04 |
| Phi 3 mini 128K instruct | Microsoft | 58.5 | 2024-07 |
| Phi 3 mini 4K instruct | Microsoft | 58.2 | 2024-07 |
| Yi 1.5 9B Chat | 01.AI | 57.8 | 2024-07 |
| Yi 1.5 9B Chat 16K | 01.AI | 57.8 | 2024-07 |
| Yi 1.5 9B | 01.AI | 56.7 | 2024-07 |
| Yi 1.5 9B 32K | 01.AI | 56.7 | 2024-07 |
| Yi 9B | 01.AI | 56.4 | 2024-07 |
| Mixtral 8x7B Instruct v0.1 | Mistral AI | 55.3 | 2026-04 |
| Yi 1.5 6B | 01.AI | 54.3 | 2024-07 |
| Meta Llama 3 8B Instruct | Meta | 52.8 | 2024-07 |
| Meta Llama 3 8B Instruct | Nous Research | 52.8 | 2024-07 |
| Nous Hermes 2 Mixtral 8x7B DPO | Nous Research | 52.8 | 2024-07 |
| Hermes 2 Theta Llama 3 8B | Nous Research | 52.1 | 2024-07 |
| Yi 1.5 6B Chat | 01.AI | 52.1 | 2024-07 |
| Mixtral 8x7B v0.1 | Mistral AI | 51.8 | 2026-04 |
| Nous Hermes 2 SOLAR 10.7B | Nous Research | 51.8 | 2024-07 |
| Mistral 7B v0.3 | Mistral AI | 51.1 | 2024-07 |
| mistral 7B v0.3 bnb 4bit | Unsloth | 51.1 | 2024-07 |
| Hermes 2 Pro Llama 3 8B | Nous Research | 50.7 | 2024-07 |
| gemma 7B it | Google DeepMind | 49.3 | 2024-07 |
| Meta Llama 3 8B | Meta | 49.3 | 2024-07 |
| Meta Llama 3 8B | Nous Research | 49.3 | 2024-07 |
| Yi 6B | 01.AI | 47.9 | 2024-07 |
| Yi 6B Chat | 01.AI | 47.9 | 2024-07 |
| Mistral 7B Instruct v0.2 | Mistral AI | 45.0 | 2024-07 |
| phi 2 | Microsoft | 44.0 | 2024-07 |
| falcon 40B | TII | 43.6 | 2024-07 |
| Qwen2 1.5B Instruct | Alibaba / Qwen Team | 43.6 | 2024-07 |
| Qwen2 0.5B Instruct | Alibaba / Qwen Team | 37.2 | 2024-07 |
| deepseek llm 7B base | DeepSeek | 36.5 | 2024-07 |
| deepseek llm 7B chat | DeepSeek | 36.5 | 2024-07 |
| deepseek coder 6.7B base | DeepSeek | 34.8 | 2024-07 |
| gemma 2B | Google DeepMind | 34.0 | 2024-07 |
| chatglm2 6B | Zhipu AI | 33.3 | 2024-07 |
| deepseek coder 6.7B instruct | DeepSeek | 31.9 | 2024-07 |
| gemma 2B it | Google DeepMind | 30.5 | 2024-07 |
| deepseek coder 1.3B base | DeepSeek | 24.5 | 2024-07 |
| deepseek coder 1.3B instruct | DeepSeek | 24.5 | 2024-07 |
| OLMo 1B hf | Allen AI | 22.0 | 2024-07 |