Llama-3.2-3B: benchmark scores, price and ranking

Llama-3.2-3B from meta-llama ranks #679 of 743 LLMs on the LLMs Tiger aggregate benchmark score (43.4), based on 1 benchmark results captured up to 2026-10-13.

Organisation: meta-llama
Type: Open-weight / local model
Open weights: Yes
Cheapest API price: $0.040 in / $0.080 out per million tokens (via llamagate, observed 2026-10-13T20:47Z)
Aggregate rank: #679 of 743 (score 43.4)

Across the 1 benchmark where at least ten models are scored, Llama-3.2-3B is in the top 10% on 0 and in the top half on 0.

Its cheapest listed input price, $0.040 per million tokens, is lower than that of 94% of the 427 models with a listed price.

Llama-3.2-3B benchmark results

Llama-3.2-3B benchmark scores, ranked against all models measured on each benchmark
Benchmark	Score	Rank	Source	Captured
MMLU-Pro	22.2%	#127 of 136	HuggingFace MMLU-Pro leaderboard	2026-10-13

About Llama-3.2-3B

Llama-3.2-3B is an open-weight model that can run on your own hardware from meta-llama. The cheapest API listing we track charges $0.040 per million input tokens and $0.080 per million output tokens, through llamagate. We hold 1 benchmark result for it across 1 area, and none is in the top tenth of its board.

Where Llama-3.2-3B stands

On our aggregate score, Llama-3.2-3B is #679 of 743 ranked models, in the lower half. Its score is 43.4; the scale is centred on 50, so it sits 6.6 points below the typical model. That beats 9% of the ranked field. Among the 3 ranked meta-llama models it is #3. The next one up is Meta-Llama-3-8B at 45.2.

Llama-3.2-3B, benchmark by benchmark

MMLU-Pro is a weak spot. Llama-3.2-3B scores 22.2%, #127 of 136, behind 93% of the models measured. The leader, Darwin-180B-RSI, reaches 88.1%. It covers 57-subject knowledge + reasoning. The board median is 55.3%, so Llama-3.2-3B is 33.1 points below it. Source: HuggingFace MMLU-Pro leaderboard, captured 2026-10-13.

Price and value

The median input price across the 427 priced models is $0.500 per million tokens, so Llama-3.2-3B costs less than the typical listing. Models that score within 1.5 points or better and cost less: Qwen2.5-Coder-7B-Instruct ($0.010, 49.5), DeepSeek V4.1 Flash (high) ($0.010, 48.9), Qwen2.5-Coder-3B-Instruct ($0.010, 46.5). Similar scores cost more on Mistral-7B-Instruct-v0.2 ($0.050) and qwen-turbo-2024-11-01 ($0.050).

Reading these numbers

The 1 score on this page comes from 1 source: HuggingFace MMLU-Pro leaderboard (1). No official leaderboard supplied any of them, so treat them as third-party measurements. With 1 results the aggregate rank can move a lot when one new score arrives.

Models ranked nearby

#676 Qwen2-1.5B (43.4)
#677 grok-2-1212 (43.4)
#678 qwen3-8b (43.4)
#680 mistral-large-2407 (43.3)
#681 granite-3.0-2b-base (43.3)
#682 open-mixtral-8x7b (43.2)