BenchLeader
Menu
Best LLMsChoose a modelLeaderboardsModelsBenchmarksProvidersReleases9
MetaOpen weights: meta-llama/Llama-2-70bCompare with another model

Llama 2-70B

Llama 2-70B is a Meta open-weights model, released 18 Jul 2023. Its best configuration ranks #772 of 785 on the BenchLeader Index at 34.3 ±2.7, in the lower half. The ± is the point: 90 other configurations score within that range, so they and this one cannot be told apart on quality alone — price and speed are what separate them. It scores highest in composite (35) and lowest in reasoning (24). Last measured 22 Oct 2026.

Blended price
–
Output speed
–
First answer
–
Context
4k
Full answer
–
Cost per run
–
Released
18 Jul 2023
How it scores by categoryDashed line = average model (50). One step of 15 = one standard deviation.
  1. Overall index34
  2. Reasoning24
  3. Coding27
  4. Maths26
  5. Human preference30
  6. Composite35

Versions

Meta has shipped 4 models under this name. Each is ranked on its own results; a newer version often has fewer results so far, which holds its index nearer the average until more arrive.

ModelReleasedIndexRank
Llama 3.3 70B6 Dec 202440.7#651
Llama 3.1 70B23 Jul 202440.9#645
Llama 3-70B18 Apr 202438.6#707
Llama 2-70Bthis page18 Jul 202334.3#772

How its index has moved

8 Sept 2026 to 23 Oct 2026
323538

The index is recomputed from scratch every day, so a line moves when a new benchmark result lands, when a publisher revises a score, or when the models it is normalised against change. Early movement usually means the score is still settling.

Benchmark results

See all LLM benchmarks. One column per reasoning effort. Rank is among every configuration of every model on that benchmark. Hover a score for the run it came from.

Reasoning

Benchmarknot statedSourceTrend
GPQA Diamond26.3%#281Epoch AI Benchmarking Hub
Arena Hard Prompts1160#332Arena (LMArena)
GPQA Diamond (AA)not in index32.7%#508Artificial Analysis
Humanity's Last Exam (AA)not in index5.2%#441Artificial Analysis
DTBenchnot in index41.6%#229Epoch AI Benchmarking Hub
ForecastBenchnot in index50.4%#84ForecastBench
HellaSwagnot in index85.3%#9Epoch AI Benchmarking Hub
BIG-Bench Hardnot in index64.9%#13Epoch AI Benchmarking Hub
WinoGrandenot in index80.2%#17Epoch AI Benchmarking Hub
ARC (AI2 Challenge)not in index78.3%#19Epoch AI Benchmarking Hub
PIQAnot in index82.8%#16Epoch AI Benchmarking Hub
CommonsenseQA 2not in index50.0%#6Epoch AI Benchmarking Hub

Coding

Benchmarknot statedSourceTrend
Arena Coding1179#334Arena (LMArena)

Maths

Benchmarknot statedSourceTrend
OTIS Mock AIME0.0%#276Epoch AI Benchmarking Hub
MATH Level 53.3%#96Epoch AI Benchmarking Hub
Arena Maths1137#323Arena (LMArena)
GSM8Knot in index69.6%#22Epoch AI Benchmarking Hub

Knowledge

Benchmarknot statedSourceTrend
MMLUnot in index69.9%#55Epoch AI Benchmarking Hub
TriviaQAnot in index87.6%#1Epoch AI Benchmarking Hub
OpenBookQAnot in index60.2%#13Epoch AI Benchmarking Hub
LAMBADAnot in index78.9%#4Epoch AI Benchmarking Hub

Instruction following

Benchmarknot statedSourceTrend
Arena Instruction Followingnot in index1135#331Arena (LMArena)

Human preference

Benchmarknot statedSourceTrend
Arena Text1171#328Arena (LMArena)
Arena Creative Writingnot in index1111#338Arena (LMArena)
Arena Multi-turnnot in index1135#326Arena (LMArena)
Arena Longer Queriesnot in index1139#334Arena (LMArena)

Composite

Benchmarknot statedSourceTrend
Epoch Capabilities Indexnot in index113.7#194Epoch AI Benchmarking Hub
AA Intelligence Index v4.3.25.3#586Artificial Analysis

More from Meta

New this month

See also

Best open weightsAll Meta models

Questions people ask

What are Llama 2-70B's benchmark scores?
As of 25 Oct 2026, Llama 2-70B scores 34.3 ±2.7 on the BenchLeader Index, strongest in composite (35) and weakest in reasoning (24). That rests on 8 independent benchmark results, and every score on this page links to its publisher.
How does Llama 2-70B rank?
Llama 2-70B is #772 of 785 ranked configurations on 25 Oct 2026. 90 other configurations score within its margin, so the exact position matters less than the band. The leader is Claude Opus 5.5 (max effort) at 72.6.
When was Llama 2-70B released?
Meta released Llama 2-70B on 18 Jul 2023.

Data as of 25 Oct 2026. Compare with another model.

Cite as: BenchLeader, “Llama 2-70B: benchmarks, pricing, speed and rank”, data as of 25 Oct 2026, https://www.benchleader.com/models/llama-2-70b.