Better than 27% of all models
Context
N/A
Input $/1M
TBD
Output $/1M
TBD
Type
text
License
Open Source
Benchmarks
12 tested
Data updated today
About
Tested on 12 benchmarks with 24.7% average. Top scores: MMLU (72.4%), Winogrande (67.0%), IFEval (33.2%).
Capabilities
coding
5.0
#164 globally
reasoning
7.1
#176 globally
math
10.8
#217 globally
knowledge
38.4
#186 globally
language
33.2
#139 globally
general
26.7
#36 globally
Benchmark Scores
Compare AllTested on 12 benchmarks · Ranked across 6 categories
Score Distribution (all 274 models)
0255075100
▲ You are here
codingCompare coding →
Cybench
5.0—Capture-the-flag cybersecurity challenges. Tests vulnerability analysis, reverse engineering, cryptography, and exploitation skills.
reasoningCompare reasoning →
MUSR
7.1—HuggingFace MuSR (Multi-Step Reasoning). Tests multi-hop reasoning requiring chaining multiple facts together.
mathCompare math →
MATH level 5
22.6—Competition-level math from AMC, AIME, and olympiad problems. Level 5 is the hardest tier, requiring creative problem-solving.
MATH Level 5
5.7—HuggingFace evaluation of MATH Level 5 problems. Competition math requiring advanced reasoning and proof construction.
OTIS Mock AIME 2024-2025
4.2—Mock AIME (American Invitational Mathematics Exam) problems from OTIS. Tests mathematical competition performance.
Excellent (85+) Good (70-85) Average (50-70) Below (<50)
Links
Research
Documentation
Community
Source Code
BenchGecko API
llama-3-70b
Specifications
- Typetext
- ContextN/A
- ReleasedJan 2024
- LicenseOpen Source
- Statusbenchmark-only
Available On
Learn More
Share & Export
Frequently Asked Questions
Llama 3-70B is an open-source text AI model by Meta, released in January 2024. It has an average benchmark score of 34.2.