Model comparison

Llama 3.2 90B vs Llama-3.3-70B-Instruct

Llama-3.3-70B-Instruct is the stronger model overall, scoring 30.6 to 27.5 on the Noometry Index.

Last verified October 10, 2026. 6 shared benchmarks.

Summary

They share 6 benchmarks with published results for both. Llama 3.2 90B scores higher in 2 categories and Llama-3.3-70B-Instruct in 2 categories; 4 gaps are clear of the uncertainty.
The widest gap is in knowledge, where Llama-3.3-70B-Instruct leads 30.6 to 21.7.
The biggest single-benchmark swing is GPQA Diamond: 41% for Llama 3.2 90B and 47.4% for Llama-3.3-70B-Instruct.

Side by side

Llama 3.2 90B and Llama-3.3-70B-Instruct specifications
	Llama 3.2 90B	Llama-3.3-70B-Instruct
Provider	Meta	Meta
Noometry Index	27.5	30.6
Released	2024-09-24	2024-12-06
Weights	Open	Open
Context window	—	128K
Max output	—	4K
Input $ / M tokens	—	$0.10
Output $ / M tokens	—	$0.32
Results tracked	9	43

Sponsored placements are available on pages like this one. Advertise on Noometry

Category by category

Coding Not comparable

Llama 3.2 90B: —, Llama-3.3-70B-Instruct: 31.0 (#290)

Coding benchmarks
Benchmark	Llama 3.2 90B	Llama-3.3-70B-Instruct
SciCode	—	26%
WeirdML	—	14.4%
BigCodeBench Instruct	—	46.9%
LiveBench Coding	—	36.6%
LMArena Coding	—	1268
BigCodeBench Complete	—	57.5%

Agentic & Tool Use Llama 3.2 90B leads

Llama 3.2 90B: 30.0 (#80), Llama-3.3-70B-Instruct: 25.8 (#105)

Agentic & Tool Use benchmarks
Benchmark	Llama 3.2 90B	Llama-3.3-70B-Instruct
BALROG	27.3%	23%
Berkeley Function Calling Leaderboard	—	31.9%

Reasoning Llama 3.2 90B leads

Llama 3.2 90B: 21.7 (#217), Llama-3.3-70B-Instruct: 14.1 (#327)

Reasoning benchmarks
Benchmark	Llama 3.2 90B	Llama-3.3-70B-Instruct
Epoch Capabilities Index	125.5	127.33
SimpleBench	—	19.9%
CritPt	—	0%
EnigmaEval	0.4%	—
LiveBench Reasoning	—	50.8%
LMArena Hard Prompts	—	1257
DTBench	—	59.5%
LiveBench Data Analysis	—	49.5%
LMCA	—	17.5%
ForecastBench	—	58.6
LiveBench	—	50.2%

Math Llama-3.3-70B-Instruct leads

Llama 3.2 90B: 11.1 (#308), Llama-3.3-70B-Instruct: 15.3 (#298)

Math benchmarks
Benchmark	Llama 3.2 90B	Llama-3.3-70B-Instruct
OTIS Mock AIME 2024-2025	2.6%	5.1%
MATH Level 5	39.4%	41.6%
LiveBench Math	—	42.2%
LMArena Math	—	1267

Knowledge Llama-3.3-70B-Instruct leads

Llama 3.2 90B: 21.7 (#274), Llama-3.3-70B-Instruct: 30.6 (#226)

Knowledge benchmarks
Benchmark	Llama 3.2 90B	Llama-3.3-70B-Instruct
GPQA Diamond	41%	47.4%
MMLU	80.3%	86.3%
Confabulations	—	22.8%
Vectara Hallucination Rate	—	4.1%
LMArena Expert	—	1225

Multimodal Not comparable

Llama 3.2 90B: 25.4 (#124), Llama-3.3-70B-Instruct: —

Multimodal benchmarks
Benchmark	Llama 3.2 90B	Llama-3.3-70B-Instruct
LMArena Vision	1000	—
GeoBench	52%	—

Multilingual Not comparable

Llama 3.2 90B: —, Llama-3.3-70B-Instruct: 39.9 (#220)

Multilingual benchmarks
Benchmark	Llama 3.2 90B	Llama-3.3-70B-Instruct
LMArena Non-English	—	1236
LMArena Chinese	—	1217
LMArena French	—	1281
LMArena German	—	1251
LMArena Japanese	—	1150
LMArena Korean	—	1143
LMArena Russian	—	1252
LMArena Spanish	—	1270

Instruction Following Not comparable

Llama 3.2 90B: —, Llama-3.3-70B-Instruct: 71.1 (#157)

Instruction Following benchmarks
Benchmark	Llama 3.2 90B	Llama-3.3-70B-Instruct
LiveBench Instruction Following	—	82.7%
LMArena Instruction Following	—	1242

Long Context Not comparable

Llama 3.2 90B: —, Llama-3.3-70B-Instruct: 26.4 (#295)

Long Context benchmarks
Benchmark	Llama 3.2 90B	Llama-3.3-70B-Instruct
Fiction.LiveBench	—	33.3%
LMArena Longer Query	—	1256

Writing & Preference Not comparable

Llama 3.2 90B: —, Llama-3.3-70B-Instruct: 47.6 (#207)

Writing & Preference benchmarks
Benchmark	Llama 3.2 90B	Llama-3.3-70B-Instruct
LMArena Text	—	1274
LMArena Creative Writing	—	1250
LMArena Multi-Turn	—	1280
LiveBench Language	—	39.2%

Frequently asked questions

Is Llama 3.2 90B better than Llama-3.3-70B-Instruct?