Neura Intelligence Index

AI Model Benchmarks

Back to Leaderboard

GPT-5.3 Codex vs Llama 3.1 70B

OpenAI 🇺🇸

42.0

#181—

Meta 🇺🇸

42.3

#178—

Category Comparison

Benchmark-by-Benchmark

coding

Benchmark	GPT-5.3 Codex	Llama 3.1 70B
SWE-Bench Pro	56.8	—
HumanEval	—	80.5

reasoning

Benchmark	GPT-5.3 Codex	Llama 3.1 70B
BigBench-Hard	—	81.2
GPQA Diamond	—	46.7
ARC-Challenge	—	94.8

general

Benchmark	GPT-5.3 Codex	Llama 3.1 70B
IFEval	—	83.6
Chatbot Arena Elo	—	1207
MMLU-Pro	—	66.4

math

Benchmark	GPT-5.3 Codex	Llama 3.1 70B
MATH	—	68
GSM8K	—	95.1

language

Benchmark	GPT-5.3 Codex	Llama 3.1 70B
HellaSwag	—	94.4
WinoGrande	—	85.3

safety

Benchmark	GPT-5.3 Codex	Llama 3.1 70B
TruthfulQA	—	60.1

Category Winners

coding

GPT-5.3 Codex

42.0

math

Llama 3.1 70B

55.6

reasoning

Llama 3.1 70B

30.5

general

Llama 3.1 70B

45.0

language

Llama 3.1 70B

70.8

safety

Llama 3.1 70B

27.6

GPT-5.3 Codex vs Llama 3.1 70B — Benchmark Comparison | Neura Market