whatmodel.sh / Llama 2 70B Chat

Llama 2 70B Chat

+ contribute

Llama 2 70B Chat is an open-weights language model from Meta, released July 2023. Its published size is 68.98B parameters and a 4k-token context window. Licence: llama2.

Catalogue identity notes

Comparison model from the Epoch AI internal-evaluation CSVs (Epoch release date 2023-07-18). Chat checkpoint; Epoch label 'Llama-2-70b-chat-hf' is the Hugging Face-format release of the same weights. 68,976,653,312 parameters (safetensors metadata).

38.5±0.8#552 of 719 LLMsfrom 28 benchmark results across 16 sources

$ how it compares

The LLMs ranked either side of Llama 2 70B Chat, out of 828 with a score, then more from Meta.

#	model	score	where	RAM	released
642	Phi-2 (2.7B) Microsoft	39±1	local	~3.2GB	2023-12
643	Granite 3.0 2B Instruct IBM	39±1	local	~3.2GB	2024-10
644	Llama 2 70B Chat Meta	39±1	local	~32GB	2023-07
645	Code Llama 13B Instruct Meta	39±1	local	~9.8GB	2023-08
646	MistralHermes-CodePro-7B-v1 beowolx	39±1	local	~6.1GB	2023-12
	more from Meta
24	Muse Spark 1.3 Meta	77±1	api $2.00/M		2026-09
40	Muse Spark 1.2 Meta	75±1	api $2.00/M		2026-08
43	Muse Spark 1.1 Meta	75±1	api $2.00/M		2026-07

see all LLMs · find the ones that fit your machine

$ what it's good at

coding▁▁▁▃▆██▇▆▄▃▁ #421 of 458 * · 4.3% of its score · 1 test

Beats DeepSeek R1 Distill Qwen 7B, Qwen3.5 2B, DeepSeek R1 Distill Llama 8B here, though they rank higher overall. Trails Llama 3.2 1B Instruct, SmolLM2 1.7B Instruct, Mistral 7B Instruct v0.1 here, though they rank lower overall.

test	result	vs the others in its run	difficulty	counts	why
Tulu 2: CodexEval pass@10	52.1% ✎	▃▃▃·▆▃▃█▃▂▂▅ beats 8/25	very easy · avg 86%	▇▇ 4.3%	sharp · at its level

chat & writing▁▁▁▃▅▄▅█▆▆▇▃ #335 of 393 · 39% of its score · 7 tests

Beats OpenThinker3 7B, Gemma 4 E2B, Code Llama 70B Instruct here, though they rank higher overall. Trails Granite 4.0 H 350M, LFM2 350M, Gemma 3 1B here, though they rank lower overall.

test	result	vs the others in its run	difficulty	counts	why
AlpacaEval 2 length-controlled	14.7% ✎	▁▂▂▃▆▆█▆▄▄▁▁ beats 36/122	medium · avg 39%	▇▇▇▇▇ 12%	typical
Arena-Hard (v0.1, GPT-4-1106 judge)	11.6% ✎	▂▁▃▅█▅▄▃▃▅▃▁ beats 5/60	medium · avg 52%	▇▇▇ 6.8%	sharp
Polyrating rating (bias-aware re-fit of Chatbot Arena votes, mid-2024)	1108.0 ✎	▂▁▂▁▃▇█▆▇▄▂▂ beats 51/95	avg 1201	▇▇ 5.4%	blunt
LMArena (text)	1170.6 ✎	▁▁▁▂▃▃▃▆▆▇█▃ beats 49/379	avg 1268	▇▇ 5.3%	blunt
MT-Bench (GPT-4 single-answer grading, LMSYS runs)	6.8600 MT-Bench score (1-10) ✎	▆▂▂▆▆▄██▂▄▄▆ beats 20/28	avg 8.819 MT-Bench score (1-10)	▇▇ 4.7%	sharp · noisy
Arabic IFEval prompt-level (Inception)	3.0% ✎	▁·▁▃▂▁▄▄▆▄█▄ beats 2/38	medium · avg 49%	▇ 3.0%	sharp
AlpacaEval 1.0 win rate (vs text-davinci-003, GPT-4 judge)	92.7% ✎	▂▂·▆·▇▃▆█▆▄▂ beats 35/45	very easy · avg 96%	▇ 1.9%	sharp · noisy

math▁▃▃▅▆▅▄█▄▁▁▁ #374 of 398 · 8.9% of its score · 3 tests

Beats Command R (03-2024), Molmo 7B-D (0924), Mistral 7B Instruct v0.3 here, though they rank higher overall. Trails Gemma 2B (base), SmolLM2 1.7B Instruct, LFM2 350M here, though they rank lower overall.

test	result	vs the others in its run	difficulty	counts	why
GSM8K +1	59.0% ✎	▆▂▅█▅▅▃▆···▂ beats 19/25	easy · avg 82%	▇▇▇ 6.3%	sharp · at its level · several runs
MATH level 5	3.3% ✎	▃▂▅▆▄█▆▃▅▂▃▄ beats 0/92	hard · avg 32%	▇ 2.4%	sharp · too hard for it
OTIS mock AIME 2024–2025	0.0% ✎	▃▃▄▃▃▃▇▇█▃▂▅ beats 1/193	very hard · avg 8.5%	▇ 0.2%	sharp · too hard for it · noisy

other tests14 not in a task · 47% of its score

test	result	vs the others in its run	difficulty	counts	why
BIG-Bench Hard +1	49.0% ✎	▁·····▂█▂▅·▁ beats 10/25	easy · avg 77%	▇▇▇▇▇▇ 14%	sharp · at its level · several runs
MMLU +1	60.9% ✎	▁▁▃▃▄█▄▃█··▁ beats 18/25	easy · avg 74%	▇▇ 6.1%	at its level · several runs
WildBench v2 WB-Score (task macro)	20.6596 WB-Score ✎	▁·▁▁·▃█▄▇▆▅▆ beats 3/52	avg 45.199 WB-Score	▇▇ 6.1%	sharp
ARC Challenge	67.7% ✎	▆▁▁▂▂▅▇█▄▃▃▂ beats 78/97	easy · avg 86%	▇▇ 4.4%	sharp · at its level
AGIEval English	46.7% ✎	█▄▄··▄··▄█·█ beats 2/9	easy · avg 69%	▇ 2.7%	at its level
EQ-Bench v1 (60 questions, paper run)	51.6% ✎	▂··▄·▂▆▄▂█·▂ beats 12/14	medium · avg 57%	▇ 2.4%	at its level
Open Persian LLM Leaderboard v2 average (21 tasks)	32.6% ✎	▃▄▂▃█▄▃▆▃▅▂▂ beats 3/81	medium · avg 58%	▇ 2.3%	blunt · at its level
BRIDGE clinical text, zero-shot average (87 tasks)	25.1% ✎	▁▁▁▁▁▂▃▃▄▆█▃ beats 18/97	hard · avg 32%	▇ 2.1%	blunt

6 more tests

test	result	vs the others in its run	difficulty	counts	why
GPQA Diamond	26.3% ✎	▁▃▁▁▂▄▆▄▅▇█▅ beats 7/208	hard · avg 40%	▇ 1.9%	sharp · too hard for it
ARC Easy	82.2% ✎	▁▂▃▂▃▅▇█▇▆▄▆ beats 64/65	easy · avg 88%	▇ 1.3%	at its level
Open PT-LLM Leaderboard average (9 Portuguese tasks)	65.7% ✎	▃▁▁▁▁▁▁▁▂█▅▁ beats 118/276	medium · avg 73%	▇ 1.1%	at its level · noisy
Tulu 2: TyDiQA gold passage, 1-shot F1	44.4% ✎	▂··▁▃▂▃▂█▅▁▂ beats 8/25	medium · avg 61%	▇ 1.0%	blunt · at its level
EuroEval Life in the UK (English knowledge MC), test split, 5-shot, accuracy	57.5% ✎	▁▁▁▁▁▁▁▃▄█▄▃ beats 29/146	easy · avg 83%	▇ 0.8%	at its level · noisy
Vectara hallucination leaderboard, 2025 dataset (HHEM-2.3, factual consistency rate)	94.1% ✎	▁▁▂▃▄▄█▆▃▂▂▁ beats 51/140	very easy · avg 94%	▇ 0.5%	blunt · noisy

$ where to run it

llama.cpp · recommended

version

Install llama.cpp · Homebrew and winget installs name the commands llama-cli and llama-server instead.
macOS / Linux
curl -LsSf https://llama.app/install.sh | sh
Windows
irm https://llama.app/install.ps1 | iex
other ways to install ↗
Run Llama 2 70B Chat · needs ~41.4 GB · from huggingface.co/TheBloke
chat in the terminal
llama cli -hf TheBloke/Llama-2-70B-Chat-GGUF:Q4_K_M
Or start a local server and open the address it prints (http://localhost:9931)
llama serve -hf TheBloke/Llama-2-70B-Chat-GGUF:Q4_K_M

LM Studio

version

Install LM Studio · Open the app once before using the lms command.
download it (macOS / Windows / Linux) ↗
Run Llama 2 70B Chat · needs ~41.4 GB · from huggingface.co/TheBloke
open in LM Studio ↗
Or download it in the terminal
lms get https://huggingface.co/TheBloke/Llama-2-70B-Chat-GGUF@Q4_K_M -y
Then run lms ls to see its name, and lms chat followed by that name

Ollama

version

Install Ollama
download it (macOS / Windows) ↗
Linux
curl -fsSL https://ollama.com/install.sh | sh
Run Llama 2 70B Chat · needs ~41.89 GB · from ollama.com
chat in the terminal
ollama run llama2:70b-chat

Weights and documentation: huggingface.co/meta-llama ↗

$ how the score was worked out · Orca 2: Teaching Small Langua… 21%, Tulu 2 — base and instruction… 19%, AlpacaEval 1.0 and 2.0 leader… 14% and 13 more

Each result implies a score on its own, after adjusting for how hard its test is and who ran it. The score is their weighted average, with a small pull toward the middle for thinly tested models. Each test's weight is its counts column above; here it's added up by who ran the tests. How it works

source	results	share of the score	licence
Orca 2: Teaching Small Language Models How to Reason (Mitra et al., Microsoft)	6	▇▇▇▇▇▇ 21%	CC BY 4.0
Tulu 2 — base and instruction-tuned evaluations	5	▇▇▇▇▇▇ 19%	CC BY 4.0
AlpacaEval 1.0 and 2.0 leaderboards (tatsu-lab/alpaca_eval repository)	2	▇▇▇▇ 14%	Apache-2.0
Arena-Hard-Auto v0.1 leaderboard (repository README, pre-v2)	1	▇▇ 7%	Apache-2.0
WildBench v2 leaderboard data (Ai2)	1	▇▇ 6%	Apache-2.0
Polyrating: A Cost-Effective and Bias-Aware Rating System for LLM Evaluation (arXiv:2409.00696v3)	1	▇▇ 5%	CC-BY-4.0
LMArena leaderboard dataset	1	▇▇ 5%	CC BY 4.0
LMSYS Chatbot Arena leaderboard Space: MT-Bench column of the published leaderboard table	1	▇ 5%	Apache-2.0
Epoch AI — internal evaluations	3	▇ 4%	CC BY 4.0
Inception Arabic/Hindi leaderboards (AraGen-v3, HindiGen-v1, Arabic IFEval)	1	▇ 3%	Apache-2.0
EQ-Bench: An Emotional Intelligence Benchmark for Large Language Models — authors' own runs	1	▇ 2%	CC BY 4.0
Open Persian LLM Leaderboard v2 (Part DP AI and AUT NLP Lab)	1	▇ 2%	Apache-2.0
BRIDGE Medical Leaderboard (Harvard Medical School / Brigham and Women's YLab)	1	▇ 2%	MIT
Open Portuguese LLM Leaderboard (archived)	1	▇ 1%	Apache-2.0
EuroEval leaderboards (results collection), English generative leaderboard	1	▇ 1%	MIT
Vectara hallucination leaderboard (vectara/hallucination-leaderboard repository)	1	▇ 1%	Apache-2.0
prior · pull toward the middle		▇ 0%
weighted average		= 38.5