Skip to content
fedi.software

GPT-6 Luna in fedi-index

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

fedi-index · General

B 1,419 ±7

Percentile 79 · #62

People prefer / Solves tasks

1,395 / 1,455

Human votes vs benchmarks, Elo-eq

Value

+80

$0.20 per 1M tokens, median of 21 providers · vendor $0.20

On your hardware

Closed weights: cloud only.

Score decomposition

B General 1,419 [1,412–1,426] Percentile 79 · #62 · components: 9
Source Component Value Elo-eq Weight Measured Variant
Epoch AI eci 156 ±1.30 1,487 ±10 31.9% 2026-09-22 —
LMArena text/overall 1,391 ±3.33 1,392 ±3 17.0% 2026-10-02 max
LMArena text/hard_prompts 1,413 ±4.04 1,406 ±4 16.8% 2026-10-02 max
LMArena text/instruction_following 1,416 ±5.21 1,428 ±5 8.0% 2026-10-02 max
LMArena text/expert 1,450 ±8.97 1,435 ±7 7.2% 2026-10-02 max
LMArena text/longer_query 1,411 ±4.85 1,411 ±4 4.1% 2026-10-02 max
LMArena text/multi_turn 1,400 ±8.16 1,399 ±7 3.6% 2026-10-02 max
Epoch AI simpleqa_verified outlier decay ×0.61 0.41 ±0.02 1,427 ±5 2.1% 2026-09-22 max
Epoch AI gpqa_diamond decay ×0.26 0.90 ±0.02 1,460 ±9 1.6% 2026-09-22 max

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

A Coding 1,424 [1,421–1,427] Percentile 80 · #55 · components: 2 provisional
Source Component Value Elo-eq Weight Measured Variant
LMArena webdev/overall 1,579 ±4.34 1,472 ±1 60.6% 2026-10-01 max
LMArena text/coding outlier 1,444 ±6.31 1,420 ±6 26.7% 2026-10-02 max

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

B Agents 1,437 [1,432–1,442] Percentile 58 · #25 · components: 1 provisional
Source Component Value Elo-eq Weight Measured Variant
LMArena agent/overall 0.01 ±0.01 1,462 ±3 90.3% 2026-10-02 max

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

A Math 1,432 [1,421–1,443] Percentile 86 · #37 · components: 3
Source Component Value Elo-eq Weight Measured Variant
LMArena text/math 1,424 ±12 1,416 ±12 43.4% 2026-10-02 max
Epoch AI frontiermath_t1_3 decay ×0.86 0.79 ±0.02 1,492 ±5 35.9% 2026-09-22 max
Epoch AI frontiermath_t4 decay ×0.86 0.56 ±0.08 1,482 ±12 11.9% 2026-09-22 max

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

B Writing 1,386 [1,377–1,394] Percentile 59 · #111 · components: 2 provisional
Source Component Value Elo-eq Weight Measured Variant
LMArena text/creative_writing 1,365 ±6.93 1,392 ±7 59.7% 2026-10-02 max
LMArena text/longer_query 1,411 ±4.85 1,411 ±4 33.4% 2026-10-02 max

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

C Language: 繁體中文 1,382 [1,367–1,398] Percentile 49 · #114 · components: 1 provisional
Source Component Value Elo-eq Weight Measured Variant
LMArena text/chinese 1,429 ±11 1,397 ±9 92.3% 2026-10-02 max

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

B Language: English 1,365 [1,356–1,375] Percentile 52 · #131 · components: 1 provisional
Source Component Value Elo-eq Weight Measured Variant
LMArena text/english 1,389 ±5.01 1,377 ±5 93.4% 2026-10-02 max

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

C Language: Español 1,392 [1,366–1,417] Percentile 43 · #86 · components: 1 provisional
Source Component Value Elo-eq Weight Measured Variant
LMArena text/spanish 1,417 ±16 1,414 ±14 89.3% 2026-10-02 max

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

B Language: Русский 1,382 [1,368–1,395] Percentile 51 · #116 · components: 1 provisional
Source Component Value Elo-eq Weight Measured Variant
LMArena text/russian 1,388 ±8.26 1,396 ±7 92.8% 2026-10-02 max

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
The same in a table
Model 2026-10
GPT-6 Luna 1,419 B

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: GPT-6 Luna

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