Meta AI·released 2024-11-26✓ 2 sources
Llama 3.3 70B benchmark scores: 10 benchmarks tracked. 30% of its scores are independently reproduced — the rest are self-reported or unverified.
Leads on 0 of 10 benchmarks·best result 81.73 on MMLU (unverified)·3 of 10 independently reproduced
| Benchmark | Score | Evidence status | Measured |
|---|---|---|---|
| MMLU | 81.73accuracy (%) | unverified· optimizedT1 | 2024-12-06Stanford CRFM Leaderboard |
| MATH level 5 | 41.6accuracy (%) | reproduced· optimizedT1 | 2024-12-06Epoch AI |
| Fiction.LiveBench | 33.3accuracy (%) | unverifiedT1 | 2024-12-06Fiction.live leaderboard |
| DTBench | 32.45accuracy (%) | unverifiedT1 | 2024-12-06https://conceptualreasoning.ai/dtbench |
| GPQA diamond | 29.92accuracy (%) | reproduced· optimizedT1 | 2024-12-06Epoch AI |
| Balrog | 23accuracy (%) | unverifiedT1 | 2024-12-06https://balrogai.com/ |
| LMCA | 20.61accuracy (%) | unverifiedT1 | 2024-12-06https://conceptualreasoning.ai/lmca |
| WeirdML | 14.44accuracy (%) | unverifiedT1 | 2024-12-06https://htihle.github.io/weirdml.html |
| OTIS Mock AIME 2024-2025 | 5.04accuracy (%) | reproduced· optimizedT1 | 2024-12-06Epoch AI |
| SimpleBench | 3.88accuracy (%) | unverifiedT1 | 2024-12-06SimpleBench Leaderboard |
Claims drawn from cited facts, not live model generation.
This model was developed by Meta AI and released in 2024. With 70.55 billion parameters, this model sits at the frontier scale, which implies substantial headroom for capability but also markedly higher training and inference costs than smaller alternatives.
3 cited facts
This model is scored on 10 tracked benchmarks, and it holds no current top score on any of them. Its record can be trusted: the scores have been independently reproduced, meaning outside evaluation backs them up rather than leaving them as vendor claims awaiting confirmation.
3 cited facts
Its mean gap to the best recorded score across its tracked benchmarks is 60.73 points.
1 cited fact
6 facts cross-checked across data sources: 1 corroborated, 5 single-source.
Source facts, citations, and refresh stamp for this record.
Sources: epoch_benchmarks, huggingface_models, openrouter_models
Llama 3.3 70B is an AI model developed by Meta AI, released 2024-11-26. Its benchmark record below tags every score as independently reproduced, self-reported, or unverified.
Llama 3.3 70B has recorded scores on 10 benchmarks, each shown with its evidence status.
Llama 3.3 70B has recorded scores on 10 benchmarks — GPQA diamond, MATH level 5, OTIS Mock AIME 2024-2025, DTBench, WeirdML, MMLU, and 4 more. The full table above shows each score with its evidence status.
3 of 10 recorded scores (30%) are independently reproduced rather than self-reported by the lab.
Llama 3.3 70B has 10 tracked claims: 3 independently reproduced, 7 unverified.