Compare AI models

Specifications
Parameters
70B—
Context window
128k1M
API pricing
Cheapest input
$0.10DeepInfra—
Cheapest output
$0.32DeepInfra—
$36.00DeepInfra · turbo · fp8—
Benchmarks
HumanEval
88.4%86.6%
MATH
77%81.1%
MMLU-Pro
68.9%64.5%
Overview
CompanyMetaQwen
Release dateDec 6 2024Nov 15 2024
AccessOpen WeightClosed
Model detailsView modelView model

Frequently asked questions

LLaMA 3.3 leads Qwen2.5-Turbo (1M) on 2 of the 3 benchmarks they both report (HumanEval, MATH, MMLU-Pro). Qwen2.5-Turbo (1M) shipped 21 days before LLaMA 3.3, so benchmark comparisons should account for the intervening progress.

Context windows are 128k (LLaMA 3.3) vs 1M (Qwen2.5-Turbo (1M)). LLaMA 3.3 is open weight, while Qwen2.5-Turbo (1M) is closed.

On HumanEval, LLaMA 3.3 leads at 88.4% vs Qwen2.5-Turbo (1M) at 86.6%. On MATH, Qwen2.5-Turbo (1M) leads at 81.1% vs LLaMA 3.3 at 77%. On MMLU-Pro, LLaMA 3.3 leads at 68.9% vs Qwen2.5-Turbo (1M) at 64.5%.