Falcon-40B
Falcon-40B places in the top three on IndoMMLU.
20
Benchmarks with a score
0
Ranked first, out of five or more models
1
In the top three
Technology Innovation Institute Mar 2023 Falcon-40B
Open weights
90.9
#229 of 256
Epoch Capabilities Index
- Open Arabic LLM Leaderboard v1 35.5% #53 /55
- IndoMMLU 32.1% #2 /5
- MedQA 42% #65 /71
Where it ranks
Its best score on each benchmark, and where that ranks among every model with a score there. We rank each model once, at its best setting.
Knowledge and language
| Benchmark | Score | Rank | Source | |
|---|---|---|---|---|
| HellaSwag | 85.3% | #10 of 63 | Claim in Falcon2-11B Technical Report, via Epoch AI | |
| PIQA | 83% | #14 of 58 | Claim in The Falcon Series of Open Language Models, via Epoch AI | |
| LAMBADA | 77.3% | #9 of 26 | Claim in The Falcon Series of Open Language Models, via Epoch AI | |
| WinoGrande | 76.9% | #29 of 76 | Claim in Llama 2: Open Foundation and Fine-Tuned Chat Models, via Epoch AI | |
| IndoMMLU | 32.1% | #2 of 5 | Official leaderboard | |
| ARC-Challenge | 61.9% | #31 of 75 | Claim in Falcon2-11B Technical Report, via Epoch AI | |
| TriviaQA | 79.9% | #16 of 38 | Claim in Llama 2: Open Foundation and Fine-Tuned Chat Models, via Epoch AI | |
| BoolQ | 83.1% | #27 of 61 | Claim in Llama 2: Open Foundation and Fine-Tuned Chat Models, via Epoch AI | |
| OpenBookQA | 56.6% | #22 of 41 | Claim in Llama 2: Open Foundation and Fine-Tuned Chat Models, via Epoch AI | |
| MMLU | 56.9% | #49 of 85 | Claim in Falcon2-11B Technical ReportFalcon2-11B Technical Report, via Epoch AI | |
| BIG-Bench Hard | 37.1% | #36 of 45 | Claim in Qwen Technical Report, via Epoch AI | |
| CMMLU | 41.5% | #9 of 11 | Official leaderboard | |
| MuSR | 36.3% | #82 of 98 | Official leaderboard | |
| AlpacaEval 2.0 | 5.61% | #42 of 47 | Official leaderboard | |
| Open Arabic LLM Leaderboard v1 | 35.5% | #53 of 55 | Official leaderboard |
Math and science
| Benchmark | Score | Rank | Source | |
|---|---|---|---|---|
| GSM8K | 33.8% | #8 of 24 | Stanford HELM, via Epoch AI |
Professional work
| Benchmark | Score | Rank | Source | |
|---|---|---|---|---|
| MedQA | 42% | #65 of 71 | Official leaderboard |
Other
| Benchmark | Score | Rank | Source | |
|---|---|---|---|---|
| HELM Classic | 72.9 | #6 of 17 | Official leaderboard | |
| Open LLM Leaderboard v2 | 11.4 | #84 of 99 | Official leaderboard | |
| HELM Lite | 21.7 | #67 of 71 | Official leaderboard |
Technology Innovation Institute's models over time
Technology Innovation Institute's models in color, against every other model on the Epoch Capabilities Index, by release date.
Source: Epoch Capabilities Index, Epoch AI (CC BY 4.0)