Llama-3.1-70B-Instruct

Meta Llama
Arena score
1261
Among open models
#100of 148
Among all on Arena
#295
Votes
55,240

Measured on Llama 3.1 70B Instruct, the served version, not a local build. LMArena, CC BY 4.0, published Nov 11, 2026.

Score by category

CategoryScore
  • English1294
  • Legal & government1283
  • Science1276
  • Hard prompts, English1261
  • French1261
  • Software & IT1260
  • Coding1260
  • Medicine & health1257
  • Multi-turn1256
  • Applied math1254
  • Spanish1253
  • Math1252
  • Hard prompts1241
  • Long prompts1241
  • Writing & literature1236
  • Business & finance1235
  • Russian1234
  • Creative writing1232
  • Instructions1231
  • Media & sports1230
  • German1222
  • Non-English1219
  • Chinese1215
  • Expert1208
  • Excluding ties1177
  • Korean1141
  • Japanese1133

Score by size

Plot against
Best for its size
ModelParamsScore
Llama-3.1-8B8.0B—
Llama-3.1-8B-Instruct8.0B1187
Llama-3.1-70B-Instruct70.6B1261
Llama-3.1-405B-Instruct406B1284

About as strong, smaller

Compare

Discussion

Built on it

Quantizations
132
Fine-tunes
109
Adapters
105
Merges
22

Ways to run it

Most downloaded

BenchClub runs

Not runnable yet

Runs open once a Llama-3.1-70B-Instruct build is registered.

More from Meta Llama

See more

Other 40–80B models

See more