MMODELYST
Models/Llama 3.2 Instruct 1B
MDL

Llama 3.2 Instruct 1B

Meta

Verified against Artificial Analysis · Jul 27, 2026
Capability
16
percentile index
Price /M
$0 / $0
in / out
Context
Avg score
9.6
11 benchmarks
AA Intelligence
1.1
index
Weights
Open
Released
Sep 25, 2024
About

Llama 3.2 Instruct 1B is a open-weights model in the Llama family from Meta. Benchmarked on 11 evals, averaging 9.6.

Benchmark scores · 11
avg 9.6 — higher is better
MATH
27.7
IFBench
23.5
GPQA Diamond
20.8
MMLU-Pro
20
Humanity's Last Exam
5.3
AA Long-Context Reasoning
5
LiveCodeBench
1.9
SciCode
1.7
AIME
0
τ²-bench
0
Terminal-Bench Hard
0
Details
PublisherMeta
FamilyLlama
Parameters
License
WeightsOpen
ReleasedSep 25, 2024
Capability over time
GPQA Diamond17.7 → 20.8
yesterday
MATH26.6 → 27.7
yesterday
IFBench22.7 → 23.5
yesterday
GPQA Diamond19.6 → 17.7
8d ago
MATH14 → 26.6
8d ago
IFBench22.8 → 22.7
8d ago

Append-only ledger — every observed change to this model's numbers.

API providers
Artificial Analysis$0/M