Skip to content
fedi.software

Llama 2 70B Chat in fedi-index

The score in every slice, what it is made of, value, local run and history.

fedi-index · General

D 1,112 ±9

Percentile 14 · #254

People prefer / Solves tasks

1,111 / 1,115

Human votes vs benchmarks, Elo-eq

Value

-267

$1.18 per 1M tokens, median of 3 providers

On your hardware

Closed weights: cloud only.

Score decomposition

D General 1,112 [1,103–1,121] Percentile 14 · #254 · components: 8
Source Component Value Elo-eq Weight Measured Variant
LMArena text/overall 1,115 ±2.78 1,115 ±3 20.7% 2026-11-11 —
LMArena text/hard_prompts 1,073 ±4.14 1,100 ±4 20.2% 2026-11-11 —
Epoch AI eci 106 ±2.89 1,089 ±23 19.4% 2023-07-18 —
LMArena text/instruction_following 1,071 ±3.89 1,101 ±4 10.1% 2026-11-11 —
LMArena text/expert 1,039 ±8.77 1,099 ±7 8.7% 2026-11-11 —
LMArena text/multi_turn 1,088 ±5.33 1,116 ±5 4.9% 2026-11-11 —
LMArena text/longer_query 1,062 ±6.44 1,085 ±6 4.6% 2026-11-11 —
Epoch AI gpqa_diamond decay ×0.25 0.26 ±0.02 1,162 ±9 2.0% 2025-01-27 —

Weight is the share of the component in the score (the rest is the prior that pulls models with little data towards the population median). Methodology

D Coding 1,108 [1,100–1,115] Percentile 10 · #258 · components: 1 provisional
Source Component Value Elo-eq Weight Measured Variant
LMArena text/coding 1,079 ±5.00 1,095 ±4 88.0% 2026-11-11 —

Weight is the share of the component in the score (the rest is the prior that pulls models with little data towards the population median). Methodology

D Math 1,110 [1,101–1,118] Percentile 12 · #226 · components: 1 provisional
Source Component Value Elo-eq Weight Measured Variant
LMArena text/math 1,091 ±5.24 1,097 ±5 87.8% 2026-11-11 —

Weight is the share of the component in the score (the rest is the prior that pulls models with little data towards the population median). Methodology

D Writing 1,100 [1,092–1,107] Percentile 11 · #234 · components: 2 provisional
Source Component Value Elo-eq Weight Measured Variant
LMArena text/creative_writing 1,075 ±5.12 1,096 ±5 62.9% 2026-11-11 —
LMArena text/longer_query 1,062 ±6.44 1,085 ±6 30.4% 2026-11-11 —

Weight is the share of the component in the score (the rest is the prior that pulls models with little data towards the population median). Methodology

D Language: 繁體中文 1,062 [1,052–1,072] Percentile 3 · #212 · components: 1 provisional
Source Component Value Elo-eq Weight Measured Variant
LMArena text/chinese 995 ±6.82 1,052 ±5 93.4% 2026-11-11 —

Weight is the share of the component in the score (the rest is the prior that pulls models with little data towards the population median). Methodology

D Language: Deutsch 1,102 [1,084–1,120] Percentile 3 · #141 · components: 1 provisional
Source Component Value Elo-eq Weight Measured Variant
LMArena text/german 1,041 ±11 1,093 ±10 91.8% 2026-11-11 —

Weight is the share of the component in the score (the rest is the prior that pulls models with little data towards the population median). Methodology

D Language: English 1,135 [1,128–1,141] Percentile 19 · #219 · components: 1 provisional
Source Component Value Elo-eq Weight Measured Variant
LMArena text/english 1,156 ±3.28 1,130 ±4 93.8% 2026-11-11 —

Weight is the share of the component in the score (the rest is the prior that pulls models with little data towards the population median). Methodology

D Language: Español 1,165 [1,140–1,189] Percentile 3 · #146 · components: 1 provisional
Source Component Value Elo-eq Weight Measured Variant
LMArena text/spanish 1,142 ±15 1,160 ±14 89.7% 2026-11-11 —

Weight is the share of the component in the score (the rest is the prior that pulls models with little data towards the population median). Methodology

D Language: Français 1,112 [1,089–1,136] Percentile 1 · #140 · components: 1 provisional
Source Component Value Elo-eq Weight Measured Variant
LMArena text/french 1,090 ±14 1,102 ±13 90.1% 2026-11-11 —

Weight is the share of the component in the score (the rest is the prior that pulls models with little data towards the population median). Methodology

D Language: 日本語 1,086 [1,063–1,108] Percentile 0 · #127 · components: 1 provisional
Source Component Value Elo-eq Weight Measured Variant
LMArena text/japanese 927 ±17 1,073 ±13 90.3% 2026-11-11 —

Weight is the share of the component in the score (the rest is the prior that pulls models with little data towards the population median). Methodology

D Language: Русский 1,127 [1,115–1,139] Percentile 7 · #215 · components: 1 provisional
Source Component Value Elo-eq Weight Measured Variant
LMArena text/russian 1,083 ±7.38 1,122 ±7 93.0% 2026-11-11 —

Weight is the share of the component in the score (the rest is the prior that pulls models with little data towards the population median). Methodology

History

fedi-index by month
1,110 1,115 1,120 1,125 2026-10 2026-11 Llama 2 70B Chat · 2026-10 · 1,124Llama 2 70B Chat · 2026-11 · 1,112 Llama 2 70B Chat
The same in a table
Model 2026-102026-11
Llama 2 70B Chat 1,1241,112 D

Badge

Show the model's fedi-index on your site or in a README. The badge updates itself; it links to this page.

fedi-index: Llama 2 70B Chat

Data CC BY 4.0: keep the link to fedi.software.