[whatmodel.sh](../../README.md) / Llama 3.3 70B Instruct
# Llama 3.3 70B Instruct
Meta · released 2024-12 · 70B params · llama-3.3
**score 55.1** ±0.5 · #280 of 732 · from 60 benchmark results across 27 sources
Score within the [LMArena (text)](https://whatmodel.sh/?set=lmarena-text&score=set) comparison set only.
## What it's good at
● marks Llama 3.3 70B Instruct among every model on that task or test, worst to best.
task vs the others rank counts tests
agent tasks ▁▁▃●█▄▅▅▄▄▂▁ #259 of 313 11% 10
coding ▁▁▁▃▅▇●▇▅▄▃▁ #243 of 469 29% 14
chat & writing ▁▁▁▃▆▄▆●▆▇▇▃ #214 of 402 9.1% 6
math ▁▁▃▅●▆▄█▄▂▁▁ #289 of 431 13% 7
web search ▁▁▁●▄█▃▄·▁·▁ #59 of 84* 1.0% 1
agent tasks · #259 of 313 · 11% of its score
Beats [DeepSeek R1 0528 Qwen3 8B](deepseek-r1-0528-qwen3-8b.md), [Magistral Small 1.0 (2506)](magistral-small-2506.md) here, though they rank higher overall. Trails [Granite 4.0 H 1B](granite-4-0-h-1b.md), [Granite 4.0 1B](granite-4-0-1b.md), [SmolLM3 3B](smollm3-3b.md) here, though they rank lower overall.
test result vs the others in its run difficulty counts why
SWE-bench Verified 29.4% ▃···▃·▃·▃●·▃ beats 4/7 hard · avg 17% 3.2% sharp · at its level
ContractBench success rate (33 ob… 7.1% ●··▆▄▂▄▂▆▆█▄ beats 1/23 very hard · avg 13% 1.7% sharp
BFCL V4 (December 2025 archive) 31.9% ▁▁▁▂█●▅▅▆▂▂▂ beats 34/77 hard · avg 31% 1.3% at its level
Finance Agent Benchmark v1 (537 q… 2.8% ●·▃▃·█▅▅▅▅▃▅ beats 1/18 very hard · avg 5.5% 1.0% sharp
EEGAgentBench overall (tool-based… 42.9% ▂·▁▁··●▇█▂·▁ beats 5/25 hard · avg 28% 0.8% blunt · at its level
τ-bench Airline 35.3% ▁▁▁·▅●▁█▇▁▁▇ beats 7/28 hard · avg 22% 0.8% at its level
BFCL v3 52.2% ▁▁▁▁▁▃▅●▅▄▁▁ beats 44/114 medium · avg 51% 0.6% blunt · at its level
τ-bench Retail 6.2% ●····▃·▂▇█▂▆ beats 0/28 hard · avg 27% 0.5% at its level · noisy
Galileo Agent Leaderboard v2 (act… 20.0% ▅●·▃▃▅▅█▃█▅▅ beats 2/19 hard · avg 26% 0.5% at its level
Galileo Agent Leaderboard v1 (too… 77.4% ▃▃▃●▆▅▆▅█▂▅▂ beats 7/31 easy · avg 78% 0.5% typical
coding · #243 of 469 · 29% of its score
Beats [Qwen3-VL 32B Instruct](qwen3-vl-32b-instruct.md), [Solar Open 100B (102B-A12B)](solar-open-100b.md), [GLM-4.7-Flash](glm-4-7-flash.md) here, though they rank higher overall. Trails [QwQ-32B-Preview](qwq-32b-preview.md), [InternLM2.5 20B Chat](internlm2-5-20b-chat.md), [Ministral 3 3B Reasoning 2512](ministral-3-3b-reasoning.md) here, though they rank lower overall.
test result vs the others in its run difficulty counts why
LiveCodeBench official (Jan 1–Nov… 31.8% ▁▁·▁▂▄▆●▅▂·▂ beats 15/31 hard · avg 23% 6.8% sharp · at its level
Aider code editing (133 Exercism … 59.4% ▁·▁▂▄▁▆●█▆▂▂ beats 29/62 medium · avg 51% 5.1% sharp · at its level
SWE-bench Verified 29.4% ▃···▃·▃·▃●·▃ beats 4/7 hard · avg 17% 3.2% sharp · at its level
Aider polyglot (225 exercises, tw… 8.4% █▄··●·█····▄ beats 4/7 very hard · avg 5.5% 2.8% sharp · noisy
HumanEval+ 77.9% ▁·▁▁▂▃▄▄█●▄▃ beats 103/145 easy · avg 72% 2.4% sharp
LiveCodeBench v5 (Oct 2024–Feb 20… 32.9% ▄▄·▄·····█▄● beats 6/7 hard · avg 18% 1.8% typical
WebApp1K-React pass@1 (onekq-ai) 33.4% ▃▁▁●▁▄▅▃▄▇▁▄ beats 9/31 medium · avg 40% 1.3% sharp · at its level · noisy
MBPP+ 81.0% ██·█●···████ beats 3/7 easy · avg 78% 1.2% typical
CodeReviewQA: change localisation… 66.6% █▂▁▁▂▂▂▃▅▅▅● beats 45/47 hard · avg 33% 1.1% sharp · at its level · noisy
BIRD-SQL (dev) 58.0% ▄··▄·▄·█··●█ beats 5/7 medium · avg 47% 1.0% at its level
CodeReviewQA: automated code refi… 32.9% ▁▁▁▁▁▁▂▂▃●▅▄ beats 29/47 hard · avg 32% 0.8% at its level
HumanEval (simple-evals) 78.9% ▂●▆█▅▃··▂··▂ beats 1/19 easy · avg 76% 0.8% typical
CodeReviewQA: solution identifica… 78.0% ▃··▁▂▂▃▂█▃●▂ beats 43/47 medium · avg 51% 0.8% sharp · at its level · noisy
CodeReviewQA: change type recogni… 66.0% ▁▁▁······▁●█ beats 14/47 easy · avg 67% 0.0% blunt · at its level · noisy
chat & writing · #214 of 402 · 9.1% of its score
Beats [Gemma 2 9B IT SimPO](gemma-2-9b-simpo.md), [Granite 4.2 8B](granite-4-2-8b.md), [Mistral Small 4 (119B-A6.5B)](mistral-small-4.md) here, though they rank higher overall. Trails [LFM2.5 1.2B Instruct](lfm2-5-1-2b-instruct.md), [Granite 4.0 H Micro 3B](granite-4-0-h-micro.md), [Gemma 3n E4B](gemma-3n-e4b.md) here, though they rank lower overall.
test result vs the others in its run difficulty counts why
LMArena (text) 1318.0 ▁▁▁▂▃▃▃●▆▇█▃ beats 150/385 avg 1260 2.9% blunt
Arabic IFEval prompt-level (Incep… 58.2% ▁·▁▁▂▁▄▄▆▄●▄ beats 32/38 medium · avg 46% 2.3% sharp · at its level
AlpacaEval 2 length-controlled 36.5% ▃▃·▁▁█●▅·▁·▁ beats 12/22 medium · avg 39% 1.9% sharp · at its level
Arena-Hard (version/judge unspeci… 65.5% ▃▃▂▅▃▆●█▅▅▂▃ beats 16/30 medium · avg 65% 1.2% at its level
IFEval prompt-level loose 90.8% ▁▃▁▃▁▄▁▁█▃▃● beats 22/22 easy · avg 78% 0.5% typical
IFEval (aggregation unspecified) 89.3% ▁▂▄▂▂▄▆█●▅▂▁ beats 109/145 easy · avg 76% 0.4% typical
math · #289 of 431 · 13% of its score
Beats [Molmo 2 8B](molmo-2-8b.md) here, though they rank higher overall. Trails [DeepSeek R1 Distill Qwen 1.5B](deepseek-r1-distill-qwen-1-5b.md), [Gemma 7B (base)](gemma-7b-base.md), [Mistral 7B v0.1 (base)](mistral-7b-v0-1.md) here, though they rank lower overall.
test result vs the others in its run difficulty counts why
MATH level 5 41.6% ▂▂▅▆●█▆▃▅▂▃▄ beats 34/92 hard · avg 28% 6.4% sharp · at its level
AIME 2024 20.0% ▁▁▁▂▁●▁▅███▃ beats 12/85 very hard · avg 14% 2.3% sharp · small run
OTIS mock AIME 2024–2025 5.1% ▃▃●▃▃▃▇▇█▃▂▅ beats 32/197 very hard · avg 7.2% 1.5% sharp · noisy
MATH (subset unspecified) 71.8% ▃▃█▂▂●·····▁ beats 18/22 medium · avg 58% 0.9% at its level
GSM8K 93.6% ▂▃··▃▂▅█▆▃●▃ beats 20/22 very easy · avg 88% 0.9% sharp · maxed out
MGSM (simple-evals) 89.1% ▁·▁▁·▁▇●█▁▃▁ beats 10/20 easy · avg 81% 0.7% typical
MathConstraint full release, no t… 8.5% ●··▄▄·▄▄█▄·▄ beats 0/8 very hard · avg 11% 0.4% typical
web search · #59 of 84 · 1.0% of its score
Beats [Kimi K2.6](kimi-k2-6.md), [GPT-5 Pro](gpt-5-pro.md), [DeepSeek V4 Flash (Preview, Apr 2026)](deepseek-v4-flash-preview.md) here, though they rank higher overall. Trails [GPT-4o mini (2024-07-18)](gpt-4o-mini-2024-07-18.md), [Mistral Small 3.1 24B (2503)](mistral-small-3-1.md), [Llama 4 Scout (109B-A17B)](llama-4-scout.md) here, though they rank lower overall.
test result vs the others in its run difficulty counts why
Finance Agent Benchmark v1 (537 q… 2.8% ●·▃▃·█▅▅▅▅▃▅ beats 1/18 very hard · avg 5.5% 1.0% sharp
other tests (24)
test result vs the others in its run difficulty counts why
Medmarks-V mean win rate (19 veri… 0.5364 ▁▄▃▃▇▄█●▅▁▁▁ beats 22/30 avg 0.418 5.4% sharp
GPQA Diamond 47.4% ▁▁▁▁▂▄●▄▅▇█▅ beats 59/211 hard · avg 39% 4.3% sharp · at its level
MMLU 85.9% ▁▂·▁·▁▁█·▃▃● beats 21/22 easy · avg 75% 3.3% typical
BIG-Bench Hard 85.8% ▁▄·▁▁·▅█·▄●▁ beats 21/22 medium · avg 72% 3.0% sharp
GPQA Diamond (simple-evals) 49.1% ▁·····▄●▂▂▃▃ beats 7/22 hard · avg 39% 2.5% sharp · at its level
Medmarks-V weighted mean win rate… 53.6% ▂▃▅▂▆▃█▅●▅▃▆ beats 22/34 medium · avg 41% 2.4% blunt · at its level
Polish medical board exams, bias-… 65.8% ▄·▂▂▂··█▂●·▄ beats 11/13 hard · avg 46% 2.0% at its level
CPTUB average (SpeakLeash, 0-5 ju… 72.9% ▁▂▂▂▃█▃●▅▄▂▃ beats 62/95 medium · avg 65% 1.9% at its level
MMLU-Pro 64.4% ▁▁·▁▁▁▃▂●▅█▁ beats 72/188 medium · avg 58% 1.8% at its level
BRIDGE clinical text, zero-shot a… 39.9% ▁▁▁▁▁▂▂▃▄▆●▃ beats 72/97 hard · avg 31% 1.7% blunt · at its level
AraGen-v3 3C3H score (Inception, … 52.1% ▁▁·▁▂▂▄█▄●▄▄ beats 48/66 hard · avg 29% 1.7% at its level
RepoQA (32K, threshold 0.8) 85.6% ▄▄▄··▄●··▄·█ beats 4/7 easy · avg 80% 1.7% typical
Polish EQ-Bench v2 (SpeakLeash) 72.9% ▁··▁▁▂▂▆▅▆█● beats 63/70 medium · avg 62% 1.4% at its level
BigCodeBench (Instruct) 41.9% ▄··▄▄···●▄█▄ beats 3/7 hard · avg 35% 1.3% at its level
Open Persian LLM Leaderboard v2 a… 68.5% ▂▄▂▃█▄▃▆▃●▂▂ beats 68/81 medium · avg 57% 1.3% blunt · at its level
MIMIC-CDM full information (4 pat… 83.9% ▄·······▄█·● beats 4/4 easy · avg 71% 1.2% typical
PopQA 48.2% ▁▁▁▃▃▃▁█▁▃·● beats 21/21 hard · avg 31% 0.9% at its level
MMLU (simple-evals, zero-shot CoT) 86.3% ▁···▂▁▅●▂▂▃▁ beats 11/22 easy · avg 80% 0.9% typical
Open PT-LLM Leaderboard average (… 79.2% ▃▁▁▁▁▁▁▁▂█●▁ beats 252/276 medium · avg 72% 0.6% at its level · noisy
DiagnosticIQ industrial maintenan… 61.7% ▁··▃▁▁▁▅●▅▁▅ beats 12/24 hard · avg 51% 0.6% blunt · at its level
EuroEval Life in the UK (English … 94.4% ▁▁▁▁▁▁▁▂▄█▄● beats 137/146 easy · avg 82% 0.3% noisy
Vectara hallucination leaderboard… 96.0% ▁▁▂▃▄▄●▆▃▂▂▁ beats 87/140 very easy · avg 94% 0.1% blunt · maxed out · noisy
DROP F1, 3-shot (simple-evals) 90.2% ▅▂▆█▆·▃·▂●·▂ beats 21/23 easy · avg 79% 0.1% blunt
TruthfulQA MC2 66.1% ▃▃▂·▃▃·▅●▂▅▂ beats 16/21 medium · avg 60% 0.1% blunt · at its level
source results share of the score
Cohere 10 17% ▇▇
Epoch AI 3 12% ▇
2 OLMo 2 Furious 8 11% ▇
Microsoft Research 9 11% ▇
Medmarks: A Comprehensive Open-Source L… 2 8% ▇
LiveCodeBench official website results 1 7% ▇
Aider code editing benchmark runs 1 5% ▇
Inception Arabic/Hindi leaderboards 2 4% ▇
LMArena leaderboard dataset 1 3% ▇
CodeReviewQA: The Code Review Comprehen… 4 3% ▇
Reassessing High-Performing LLMs on Pol… 1 2% ▇
Berkeley Function Calling Leaderboard r… 2 2% ▇
SpeakLeash CPTUB 1 2% ▇
BRIDGE Medical Leaderboard 1 2% ▇
ContractBench: Can LLM Agents Preserve … 1 2% ▇
SpeakLeash Polish EQ-Bench leaderboard 1 1% ▇
WebApp1K models leaderboard 1 1% ▇
Open Persian LLM Leaderboard v2 1 1% ▇
MIMIC Clinical Decision Making leaderbo… 1 1% ▇
Galileo Agent Leaderboard 2 1% ▇
Finance Agent Benchmark 1 1% ▇
EEGAgentBench: Benchmarking LLM Agents … 1 1% ▇
Open Portuguese LLM Leaderboard 1 1% ▇
DiagnosticIQ: A Benchmark for LLM-Based… 1 1% ▇
MathConstraint: Automated Generation of… 1 0% ▇
EuroEval leaderboards 1 0% ▇
Vectara hallucination leaderboard 1 0% ▇
prior: pull toward the middle 0% ▇
weighted average = 55.1
Licences: Cohere ([CC BY 4.0](../../licenses/report-2504.00698v2.txt)) · Epoch AI ([CC BY 4.0](https://creativecommons.org/licenses/by/4.0/)) · 2 OLMo 2 Furious ([CC BY 4.0](../../licenses/paper-2501.00656v3-CC-BY-4.0.txt)) · Microsoft Research ([CC BY 4.0](../../licenses/report-2412.08905v1.txt)) · Medmarks: A Comprehensive Open-Source LLM Benchmark Suite for Medical Tasks ([CC BY 4.0](../../licenses/paper-2605.01417v1-CC-BY-4.0.txt)) · LiveCodeBench official website results ([CC-BY-SA-4.0](../../licenses/lcb-sharealike-NOTICE.txt)) · Aider code editing benchmark runs ([Apache-2.0](../../licenses/aider-polyglot-runs-Apache-2.0.txt)) · Inception Arabic/Hindi leaderboards ([Apache-2.0](../../licenses/hf-leaderboard-spaces-Apache-2.0.txt)) · LMArena leaderboard dataset ([CC BY 4.0](https://creativecommons.org/licenses/by/4.0/)) · CodeReviewQA: The Code Review Comprehension Assessment for Large Language Models ([CC-BY-4.0](https://creativecommons.org/licenses/by/4.0/)) · Reassessing High-Performing LLMs on Polish Medical Exams: True Competence or Bias-Driven Performance? ([CC BY-SA 4.0](https://creativecommons.org/licenses/by-sa/4.0/)) · Berkeley Function Calling Leaderboard result archive ([Apache-2.0](../../licenses/bfcl-result-archive-v4-Apache-2.0.txt)) · SpeakLeash CPTUB ([MIT](../../licenses/hf-leaderboard-spaces-MIT.txt)) · BRIDGE Medical Leaderboard ([MIT](../../licenses/bridge-medical-leaderboard-MIT.txt)) · ContractBench: Can LLM Agents Preserve Observation Contracts? ([CC BY 4.0](https://creativecommons.org/licenses/by/4.0/)) · SpeakLeash Polish EQ-Bench leaderboard ([Apache-2.0](../../licenses/hf-leaderboard-spaces-Apache-2.0.txt)) · WebApp1K models leaderboard ([MIT](../../licenses/webapp1k-leaderboard-MIT.txt)) · Open Persian LLM Leaderboard v2 ([Apache-2.0](../../licenses/hf-leaderboard-spaces-Apache-2.0.txt)) · MIMIC Clinical Decision Making leaderboard ([Apache-2.0](../../licenses/hf-leaderboard-spaces-Apache-2.0.txt)) · Galileo Agent Leaderboard ([Apache-2.0](../../licenses/galileo-agent-leaderboard-Apache-2.0.txt)) · Finance Agent Benchmark ([CC-BY-4.0](https://creativecommons.org/licenses/by/4.0/)) · EEGAgentBench: Benchmarking LLM Agents on Short- and Long-Horizon EEG Analysis ([CC BY 4.0](https://creativecommons.org/licenses/by/4.0/)) · Open Portuguese LLM Leaderboard ([Apache-2.0](../../licenses/hf-leaderboard-spaces-Apache-2.0.txt)) · DiagnosticIQ: A Benchmark for LLM-Based Industrial Maintenance Action Recommendation from Symbolic Rules ([CC BY 4.0](https://creativecommons.org/licenses/by/4.0/)) · MathConstraint: Automated Generation of Verified Combinatorial Reasoning Instances for LLMs ([CC BY 4.0](../../licenses/paper-2605.08498v1-CC-BY-4.0.txt)) · EuroEval leaderboards ([MIT](https://github.com/EuroEval/leaderboards/blob/1a77f1d2f8ba1ba05ab73287fb090d76bf41b2e0/LICENSE)) · Vectara hallucination leaderboard ([Apache-2.0](https://www.apache.org/licenses/LICENSE-2.0))
## How to run it
### Through an API
api (Amazon Bedrock) · $0.72/M blended · $0.72 in / $0.72 out per 1M tokens · [docs.aws.amazon.com](https://docs.aws.amazon.com/bedrock/latest/userguide/models-supported.html)
[price source](https://github.com/anomalyco/models.dev/blob/a668908cfe7073faa0363699486fc41a3ff52412/providers/amazon-bedrock/models/us.meta.llama3-3-70b-instruct-v1:0.toml): Standard base token rate from models.dev. Provider-specific tiers, long-context surcharges, caching and promotions may differ. · [context source](https://github.com/meta-llama/llama-models/blob/0e0b8c519242d5833d8c11bffc1232b77ad7f301/models/llama3_3/MODEL_CARD.md): Meta Llama 3.3 model card (llama-models repo, pinned): 70B context length 128k (stated rounded). Model ceiling; the priced Amazon Bedrock endpoint may apply its own limit. Observed 2026-09-25.
### 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 3.3 70B Instruct (ud-q2_k_xl, needs ~21.5 GB, from [huggingface.co/unsloth](https://huggingface.co/unsloth/Llama-3.3-70B-Instruct-GGUF/tree/main))
```sh
llama cli -hf unsloth/Llama-3.3-70B-Instruct-GGUF:UD-Q2_K_XL
```
3. Or start a local server and open the address it prints (http://localhost:9931)
```sh
llama serve -hf unsloth/Llama-3.3-70B-Instruct-GGUF:UD-Q2_K_XL
```
### 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 3.3 70B Instruct (4bit, needs ~39.7 GB, Mac only, from [huggingface.co/mlx-community](https://huggingface.co/mlx-community/Llama-3.3-70B-Instruct-4bit/tree/de2dfaf56839b7d0e834157d2401dee02726874d))
```sh
lms get https://huggingface.co/mlx-community/Llama-3.3-70B-Instruct-4bit -y
```
3. Then run lms ls to see its name, and lms chat followed by that name
Other versions: ud-q2_k_xl (~21.5 GB) `lms get https://huggingface.co/unsloth/Llama-3.3-70B-Instruct-GGUF@UD-Q2_K_XL -y` · 8bit (~75 GB, Mac only) `lms get https://huggingface.co/mlx-community/Llama-3.3-70B-Instruct-8bit -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 3.3 70B Instruct (Q4_K_M, needs ~42.5 GB, from [ollama.com](https://ollama.com/library/llama3.3/tags))
```sh
ollama run llama3.3:70b
```
### MLX (Mac)
1. Install MLX (Mac). Apple silicon Macs only, text only.
```sh
# with uv
uv tool install mlx-lm
# or
pip install mlx-lm
```
2. Run Llama 3.3 70B Instruct (4bit, needs ~39.7 GB, Mac only, from [huggingface.co/mlx-community](https://huggingface.co/mlx-community/Llama-3.3-70B-Instruct-4bit/tree/de2dfaf56839b7d0e834157d2401dee02726874d))
```sh
mlx_lm.chat --model mlx-community/Llama-3.3-70B-Instruct-4bit
```
3. Or start an OpenAI-compatible server at http://localhost:8080/v1
```sh
mlx_lm.server --model mlx-community/Llama-3.3-70B-Instruct-4bit
```
Other versions: 8bit (~75 GB, Mac only) `mlx_lm.chat --model mlx-community/Llama-3.3-70B-Instruct-8bit`
Model repository and documentation: [huggingface.co/meta-llama](https://huggingface.co/meta-llama/Llama-3.3-70B-Instruct)
## Published speed measurements
Reference results for the hardware, engine and settings below. These are not verified speeds of the downloads above or predictions for your machine.
hardware precision speed engine and settings
NVIDIA GeForce RTX 5090 GGUF IQ3_XXS (U… 46.5 tok/s llama.cpp (GGUF; version not stat…
NVIDIA GeForce RTX 5090 GGUF Q4_K_M (Un… 2.3 tok/s llama.cpp (GGUF; version not stat…
Apple M4 Pro (14-core CPU /… MLX 4-bit 5.1 tok/s MLX v0.30.6 (macOS 26 Tahoe) · Of…
Apple M3 Ultra (28-core CPU… MLX 4-bit 13.1 tok/s MLX v0.30.6 (macOS 26 Tahoe) · Of…
Each speed links to its source, which has the full settings.
## Quantization evidence
How much of the full-precision result a quantized version kept, measured within one study. Rows from the same study share a table and aren't independent. They don't describe every download of that precision.
precision test full → quantized kept
AutoRound 2-bit (group size 3… WMT24++ X->English COMET (wmt22-c… 0.795 → 0.767 96.5%
AutoRound 2-bit (group size 3… WMT24++ English->X COMET (wmt22-c… 0.776 → 0.667 86.0%
AutoRound 4-bit (group size 1… WMT24++ X->English COMET (wmt22-c… 0.795 → 0.795 100.0%
AutoRound 4-bit (group size 1… WMT24++ English->X COMET (wmt22-c… 0.776 → 0.773 99.6%
AWQ 4-bit (AutoAWQ, zero poin… WMT24++ X->English COMET (wmt22-c… 0.795 → 0.795 100.0%
AWQ 4-bit (AutoAWQ, zero poin… WMT24++ English->X COMET (wmt22-c… 0.776 → 0.772 99.5%
bitsandbytes NF4 (Transformer… WMT24++ English->X COMET (wmt22-c… 0.776 → 0.762 98.2%
bitsandbytes NF4 (Transformer… WMT24++ X->English COMET (wmt22-c… 0.795 → 0.768 96.6%
GGUF Q2_K with imatrix (llama… WMT24++ X->English COMET (wmt22-c… 0.795 → 0.776 97.6%
GGUF Q2_K with imatrix (llama… WMT24++ English->X COMET (wmt22-c… 0.776 → 0.743 95.8%
GGUF Q4_K_M with imatrix (lla… WMT24++ X->English COMET (wmt22-c… 0.795 → 0.793 99.7%
GGUF Q4_K_M with imatrix (lla… WMT24++ English->X COMET (wmt22-c… 0.776 → 0.772 99.6%
Each kept value links to its source.
[interactive version](https://whatmodel.sh/model/llama-3-3-70b) · [JSON](https://whatmodel.sh/api?format=json&all=1)