[whatmodel.sh](../../README.md) / Llama 2 70B (base)
# Llama 2 70B (base)
Meta · released 2023-07 · 69B params · llama2
**score 42.7** ±0.8 · #521 of 732 · from 22 benchmark results across 6 sources
Score within the [LMArena (text)](https://whatmodel.sh/?set=lmarena-text&score=set) comparison set only.
## What it's good at
● marks Llama 2 70B (base) among every model on that task or test, worst to best.
task vs the others rank counts tests
coding ▁▁▁●▅▇█▇▅▄▃▁ #420 of 469 19% 3
math ▁▁▃▅█●▄█▄▂▁▁ #232 of 431 16% 2
coding · #420 of 469 · 19% of its score
Beats [DeepSeek R1 Distill Qwen 7B](deepseek-r1-distill-qwen-7b.md), [Qwen3.5 2B](qwen3-5-2b.md), [DeepSeek R1 Distill Llama 8B](deepseek-r1-distill-llama-8b.md) here, though they rank higher overall. Trails [Llama 3.2 1B Instruct](llama-3-2-1b.md), [SmolLM2 1.7B Instruct](smollm2-1-7b.md), [Code Llama 7B (base)](codellama-7b.md) here, though they rank lower overall.
test result vs the others in its run difficulty counts why
HumanEval 29.3% █···█·██●··█ beats 4/5 medium · avg 61% 11% sharp · at its level
Tulu 2: CodexEval pass@10 60.1% ▃▃▃·▆●▃█▃▂▂▅ beats 11/25 easy · avg 83% 4.4% sharp · at its level
MBPP 49.8% ▄··▄·▄··●··▄ beats 3/5 easy · avg 69% 3.9% at its level
math · #232 of 431 · 16% of its score
Beats [GPT-4o (2024-08-06)](gpt-4o-2024-08-06.md), [Command A (03-2025)](command-a.md), [GPT-4o (2024-05-13)](gpt-4o-2024-05-13.md) here, though they rank higher overall. Trails [Gemma 7B (base)](gemma-7b-base.md), [Phi-2 (2.7B)](phi-2.md), [Granite 8B Code Base (4K)](granite-8b-code-base-4k.md) here, though they rank lower overall.
test result vs the others in its run difficulty counts why
GSM8K 55.5% ▆▂▅█▅●▃▆···▂ beats 18/25 easy · avg 82% 12% sharp · at its level · several runs
MATH 13.8% █·█·█·█●···█ beats 4/5 hard · avg 32% 3.8% typical
other tests (15)
test result vs the others in its run difficulty counts why
MMLU 64.5% ▁▁▃▃▄█▄▁●··▁ beats 19/25 easy · avg 82% 12% at its level · several runs
MMLU-Pro 36.3% ▃▂●█▃·█▃▅··▂ beats 4/23 medium · avg 54% 11% at its level
BIG-Bench Hard 66.0% ▁·····▂█▂●·▁ beats 20/25 easy · avg 74% 11% sharp · at its level
ARC Challenge 56.5% ▄···▄···█●·▄ beats 4/5 easy · avg 77% 8.0% sharp · at its level
WinoGrande 80.4% █··█·█·██··● beats 5/5 easy · avg 86% 4.2% at its level
Natural Questions (closed book) 25.4% █·····▄▄●··▄ beats 4/5 medium · avg 41% 4.2% at its level
TriviaQA 73.0% █···██··█·█● beats 5/5 easy · avg 83% 3.5% at its level
HellaSwag 85.4% █···██···██● beats 5/5 very easy · avg 91% 2.8% typical
Open PT-LLM Leaderboard average (… 67.7% ▃▁▁▁▁▁▁▁▂●▅▁ beats 139/276 medium · avg 72% 1.7% at its level · noisy
ARC Easy 79.9% ▄···▄···●▄·▄ beats 3/5 easy · avg 86% 1.5% blunt
EuroEval HellaSwag (English), tes… 59.0% ▃·▁▁▂▁▂●█▇▄▁ beats 33/100 medium · avg 67% 1.3% sharp · at its level · noisy
EuroEval Life in the UK (English … 86.0% ▁▁▁▁▁▁▁▂▄●▄▃ beats 101/146 easy · avg 82% 1.2% at its level · noisy
Tulu 2: TyDiQA gold passage, 1-sh… 62.6% ▂··▁▃▂▂▂█▅▁● beats 24/25 medium · avg 59% 1.1% blunt · at its level
PIQA 82.6% ▄····▄··█●·▄ beats 4/5 easy · avg 87% 1.1% blunt · at its level
COPA 97.6% ▅·▃·▃··█▃··● beats 8/8 easy · avg 92% 0.3% noisy
source results share of the score
Mixtral of Experts 12 58% ▇▇▇▇▇▇▇
Tulu 2 5 26% ▇▇▇
MMLU-Pro author-released evaluation runs 1 11% ▇
EuroEval leaderboards 2 2% ▇
Open Portuguese LLM Leaderboard 1 2% ▇
From Test-Taking to Test-Making: Examin… 1 0% ▇
prior: pull toward the middle 0% ▇
weighted average = 42.7
Licences: Mixtral of Experts ([CC BY 4.0](../../licenses/CC-BY-4.0.txt)) · Tulu 2 ([CC BY 4.0](../../licenses/CC-BY-4.0.txt)) · MMLU-Pro author-released evaluation runs ([Apache-2.0](../../licenses/mmlu-pro-author-runs-Apache-2.0.txt)) · EuroEval leaderboards ([MIT](https://github.com/EuroEval/leaderboards/blob/1a77f1d2f8ba1ba05ab73287fb090d76bf41b2e0/LICENSE)) · Open Portuguese LLM Leaderboard ([Apache-2.0](../../licenses/hf-leaderboard-spaces-Apache-2.0.txt)) · From Test-Taking to Test-Making: Examining LLM Authoring of Commonsense Assessment Items ([CC-BY-4.0](https://creativecommons.org/licenses/by/4.0/))
## How to run it
### llama.cpp (recommended)
1. Install llama.cpp. Homebrew and winget installs name the commands llama-cli and llama-server instead. [Other ways to install](https://github.com/ggml-org/llama.cpp/blob/master/docs/install.md)
```sh
# macOS / Linux
curl -LsSf https://llama.app/install.sh | sh
# Windows
irm https://llama.app/install.ps1 | iex
```
2. Run Llama 2 70B (base) (Q4_K_M, needs ~41.4 GB, from [huggingface.co/TheBloke](https://huggingface.co/TheBloke/Llama-2-70B-GGUF/tree/94301663ac62f467d3b2fe8deab5cc288fef7c3a))
```sh
llama cli -hf TheBloke/Llama-2-70B-GGUF:Q4_K_M
```
3. Or start a local server and open the address it prints (http://localhost:9931)
```sh
llama serve -hf TheBloke/Llama-2-70B-GGUF:Q4_K_M
```
Other versions: Q2_K (~29.3 GB) `llama cli -hf TheBloke/Llama-2-70B-GGUF:Q2_K`
### LM Studio
1. Install LM Studio. Open the app once before using the lms command. [Download it (macOS / Windows / Linux)](https://lmstudio.ai/download)
2. Run Llama 2 70B (base) (Q4_K_M, needs ~41.4 GB, from [huggingface.co/TheBloke](https://huggingface.co/TheBloke/Llama-2-70B-GGUF/tree/94301663ac62f467d3b2fe8deab5cc288fef7c3a))
```sh
lms get https://huggingface.co/TheBloke/Llama-2-70B-GGUF@Q4_K_M -y
```
3. Then run lms ls to see its name, and lms chat followed by that name
Other versions: Q2_K (~29.3 GB) `lms get https://huggingface.co/TheBloke/Llama-2-70B-GGUF@Q2_K -y`
### Ollama
1. Install Ollama. [Download it (macOS / Windows)](https://ollama.com/download)
```sh
# Linux
curl -fsSL https://ollama.com/install.sh | sh
```
2. Run Llama 2 70B (base) (Q4_0, needs ~39 GB, from [ollama.com](https://ollama.com/library/llama2/tags))
```sh
ollama run llama2:70b-text
```
Model repository and documentation: [huggingface.co/meta-llama](https://huggingface.co/meta-llama/Llama-2-70b-hf)
[interactive version](https://whatmodel.sh/model/llama-2-70b) · [JSON](https://whatmodel.sh/api?format=json&all=1)