Meta-Llama-3-8B-Instruct: benchmark scores, price and ranking

Meta-Llama-3-8B-Instruct from Meta AI ranks #738 of 743 LLMs on the LLMs Tiger aggregate benchmark score (39.4), based on 5 benchmark results captured up to 2026-10-13.

Organisation: Meta AI
Type: Open-weight / local model
Open weights: Yes
Released: 2024-04-18 (source: epoch)
Cheapest API price: $0.030 in / $0.060 out per million tokens (via deepinfra, observed 2026-10-13T20:47Z)
Aggregate rank: #738 of 743 (score 39.4)

Relative to other models, Meta-Llama-3-8B-Instruct is strongest on MMLU-Pro (#97 of 136) and weakest on DTBench (#143 of 146).

Across the 5 benchmarks where at least ten models are scored, Meta-Llama-3-8B-Instruct is in the top 10% on 0 and in the top half on 0.

Its cheapest listed input price, $0.030 per million tokens, is lower than that of 96% of the 427 models with a listed price.

Meta-Llama-3-8B-Instruct benchmark results

Meta-Llama-3-8B-Instruct benchmark scores, ranked against all models measured on each benchmark
Benchmark	Score	Rank	Source	Captured
MMLU-Pro	41.0%	#97 of 136	HuggingFace MMLU-Pro leaderboard	2026-10-13
OTIS Mock AIME	1.9%	#174 of 189	Epoch AI Benchmarking Hub	2026-10-13
Chess Puzzles	0.0%	#116 of 122	Epoch AI Benchmarking Hub	2026-10-13
GPQA Diamond	26.1%	#264 of 275	Epoch AI Benchmarking Hub	2026-10-13
DTBench	43.9%	#143 of 146	Epoch AI Benchmarking Hub	2026-10-13

About Meta-Llama-3-8B-Instruct

Meta-Llama-3-8B-Instruct is an open-weight model that can run on your own hardware from Meta AI. It was released on 2024-04-18, according to Epoch AI. The cheapest API listing we track charges $0.030 per million input tokens and $0.060 per million output tokens, through deepinfra. We hold 5 benchmark results for it across 3 areas, and none is in the top tenth of its board.

Where Meta-Llama-3-8B-Instruct stands

On our aggregate score, Meta-Llama-3-8B-Instruct is #738 of 743 ranked models, in the lower half. Its score is 39.4; the scale is centred on 50, so it sits 10.6 points below the typical model. That beats 1% of the ranked field. Among the 28 ranked Meta AI models it is #27. The next one up is Llama-4-Scout-17B-16E-Instruct at 40.1. The next one down is Llama-3.1-8B-Instruct at 36.7.

Meta-Llama-3-8B-Instruct, benchmark by benchmark

Meta-Llama-3-8B-Instruct lands below the middle on MMLU-Pro: 41.0%, #97 of 136, with 71% of the field scoring higher. Darwin-180B-RSI leads at 88.1%; the closest result above is Mixtral-8x7B-v0.1 at 41.0%. It covers 57-subject knowledge + reasoning. The board median is 55.3%, so Meta-Llama-3-8B-Instruct is 14.3 points below it. Source: HuggingFace MMLU-Pro leaderboard, captured 2026-10-13.

OTIS Mock AIME is a weak spot. Meta-Llama-3-8B-Instruct scores 1.9%, #174 of 189, behind 92% of the models measured. The leader, Claude Fable 5.1 (max), reaches 100.0%. It covers competition maths (AIME 2024-25 style). The board median is 57.8%, so Meta-Llama-3-8B-Instruct is 55.8 points below it. Source: Epoch AI Benchmarking Hub, captured 2026-10-13.

Chess Puzzles is a weak spot. Meta-Llama-3-8B-Instruct scores 0.0%, #116 of 122, behind 94% of the models measured. The leader, GPT-6 Astra (max), reaches 72.0%. It covers chess tactics puzzles. The board median is 12.5%, so Meta-Llama-3-8B-Instruct is 12.5 points below it. Source: Epoch AI Benchmarking Hub, captured 2026-10-13.

GPQA Diamond is a weak spot. Meta-Llama-3-8B-Instruct scores 26.1%, #264 of 275, behind 96% of the models measured. The leader, GPT-6 Astra (max), reaches 95.8%. It covers expert-level science questions. The board median is 71.7%, so Meta-Llama-3-8B-Instruct is 45.6 points below it. Source: Epoch AI Benchmarking Hub, captured 2026-10-13.

DTBench is a weak spot. Meta-Llama-3-8B-Instruct scores 43.9%, #143 of 146, behind 97% of the models measured. The leader, Claude Opus 5.5 (max), reaches 98.9%. It covers dynamic tool-use benchmark. The board median is 68.8%, so Meta-Llama-3-8B-Instruct is 24.9 points below it. Source: Epoch AI Benchmarking Hub, captured 2026-10-13.

Strengths and weaknesses

Reasoning, maths and science: weak, ahead of 14% of the field on average across 3 benchmarks. Its best board here is MMLU-Pro (#97 of 136) and its weakest is GPQA Diamond (#264 of 275).

Price and value

The median input price across the 427 priced models is $0.500 per million tokens, so Meta-Llama-3-8B-Instruct 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 qwen2.5-7b-instruct ($0.040) and Llama-4-Scout-17B-16E-Instruct ($0.050).

Release history

The Meta AI model released before it was Meta-Llama-3-70B-Instruct on 2024-04-18, which scores 42.5. The next one was Llama 3.1-405B on 2024-07-23, at 51.2.

Reading these numbers

The 5 scores on this page come from 2 sources: Epoch AI Benchmarking Hub (4), HuggingFace MMLU-Pro leaderboard (1). No official leaderboard supplied any of them, so treat them as third-party measurements. With 5 results the aggregate rank can move a lot when one new score arrives.

Models ranked nearby

#735 qwen2.5-7b-instruct (39.9)
#736 claude-3-haiku-20240307 (39.9)
#737 gemma-3-27b-it (39.7)
#739 gemma-3-12b-it (39.1)
#740 GPT-4.1 nano (38.7)
#741 GPT-3.5 Turbo (Jan 2024) (38.5)