Benchmarks / 3D
BuzzBench
An LLM-judged evaluation of humour explanation and predicted audience response to Never Mind the Buzzcocks introductions. Maintained by Samuel Paech / EQ-Bench. Score: LLM judge score against human-authored gold responses; headline Score column.
BuzzBench
Higher is better · best effort per model
full table below
| # | Model | Effort | Score |
|---|---|---|---|
| 1 | Gemini 2.5 Pro | high | 71 |
| 2 | GPT-4.1 | high | 67 |
| 3 | DeepSeek R1 | xhigh | 63 |
| 4 | Sonnet 3.5 | medium | 62 |
| 5 | o1 | high | 59 |
| 6 | Gemini 2.0 Flash | medium | 54 |
| 7 | Grok 2.1212 | – | 52 |
| 8 | DeepSeek V3 | none | 52 |
| 9 | Mistral Large 2411 | – | 47 |
| 10 | o1 mini | high | 46 |
| 11 | Llama 3.1 405b | none | 46 |
| 12 | Qwen 2.5 72B | none | 46 |
| 13 | Llama 3.3 70b | none | 46 |
| 14 | GPT-3.5 | high | 39 |
| 15 | Llama 3.1 8b | medium | 33 |
| 16 | Qwen 2.5 7B | – | 31 |
| 17 | Llama 3.2 3b | none | 27 |
| 18 | Llama 3.2 1b | none | 10 |