MMODELYST
Models/Llama 3 Instruct 8B
MDL

Llama 3 Instruct 8B

Meta

Verified against Artificial Analysis · Jul 13, 2026
Capability
26
percentile index
Price /M
$0.045 / $0.145
in / out
Context
Avg score
17.1
10 benchmarks
AA Intelligence
1.2
index
Speed
77 tok/s
TTFT 0.5s
Released
Apr 18, 2024
About

Llama 3 Instruct 8B is a open-weights model in the Llama family from Meta. Benchmarked on 10 evals, averaging 17.1.

Benchmark scores · 10
avg 17.1 — higher is better
MATH
49.9
MMLU-Pro
40.5
GPQA Diamond
29.6
IFBench
24.6
SciCode
11.9
LiveCodeBench
9.6
Humanity's Last Exam
5.1
τ²-bench
0
Terminal-Bench Hard
0
AA Long-Context Reasoning
0
Details
PublisherMeta
FamilyLlama
Parameters
License
WeightsOpen
ReleasedApr 18, 2024
Performance & price
Output speed77 tok/s
Latency (TTFT)0.5s
Price in / out$0.045 / $0.145 per 1M
Blended price$0.07 / 1M

Median across API providers, via Artificial Analysis. Blended = 3:1 input:output. For reasoning models, latency includes thinking time. Methodology →

Capability over time

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

API providers
Artificial Analysis$0.145/M