GPT-5.6 Luna in fedi-index
The score in every slice, what it is made of, value, local run and history.
fedi-index · General
A 1,438 ±6
Percentile 86 · #41
People prefer / Solves tasks
1,426 / 1,457
Human votes vs benchmarks, Elo-eq
Value
+80
$0.45 per 1M tokens, median of 33 providers · vendor $0.45
On your hardware
Closed weights: cloud only.
Score decomposition
A General 1,438 [1,432–1,444] Percentile 86 · #41 · components: 9
| Source | Component | Value | Elo-eq | Weight | Measured | Variant |
|---|---|---|---|---|---|---|
| Epoch AI |
eci
|
156 ±1.20 | 1,488 ±10 | 31.8% | 2026-07-09 | — |
| LMArena |
text/overall
|
1,431 ±2.23 | 1,431 ±2 | 16.6% | 2026-10-02 | xhigh |
| LMArena |
text/hard_prompts
|
1,452 ±2.60 | 1,442 ±2 | 16.5% | 2026-10-02 | xhigh |
| LMArena |
text/instruction_following
|
1,438 ±3.17 | 1,450 ±3 | 8.2% | 2026-10-02 | xhigh |
| LMArena |
text/expert
|
1,479 ±4.85 | 1,459 ±4 | 7.9% | 2026-10-02 | xhigh |
| LMArena |
text/longer_query
|
1,437 ±2.99 | 1,435 ±3 | 4.1% | 2026-10-02 | xhigh |
| LMArena |
text/multi_turn
|
1,433 ±4.23 | 1,429 ±4 | 4.0% | 2026-10-02 | xhigh |
| Epoch AI |
simpleqa_verified
outlier decay ×0.61 |
0.41 ±0.02 | 1,426 ±5 | 2.0% | 2026-08-10 | max |
| Epoch AI |
gpqa_diamond
decay ×0.26 |
0.92 ±0.02 | 1,466 ±8 | 1.5% | 2026-07-09 | 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 81 · #54 · components: 2 provisional
| Source | Component | Value | Elo-eq | Weight | Measured | Variant |
|---|---|---|---|---|---|---|
| LMArena |
webdev/overall
|
1,518 ±3.27 | 1,455 ±1 | 46.0% | 2026-10-01 | xhigh-codex-harness |
| LMArena |
text/coding
|
1,466 ±3.45 | 1,440 ±3 | 44.3% | 2026-10-02 | xhigh |
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 Agents 1,426 [1,424–1,429] Percentile 42 · #34 · components: 1 provisional
| Source | Component | Value | Elo-eq | Weight | Measured | Variant |
|---|---|---|---|---|---|---|
| LMArena |
agent/overall
|
-0.01 ±0.00 | 1,450 ±2 | 90.5% | 2026-10-02 | xhigh |
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 1,450 [1,442–1,458] Percentile 91 · #23 · components: 3
| Source | Component | Value | Elo-eq | Weight | Measured | Variant |
|---|---|---|---|---|---|---|
| LMArena |
text/math
|
1,458 ±7.25 | 1,449 ±7 | 50.2% | 2026-10-02 | xhigh |
| Epoch AI |
frontiermath_t1_3
decay ×0.86 |
0.82 ±0.02 | 1,499 ±5 | 31.7% | 2026-07-09 | max |
| Epoch AI |
frontiermath_t4
decay ×0.86 |
0.61 ±0.08 | 1,490 ±12 | 10.5% | 2026-07-09 | 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 Vision 1,424 [1,420–1,429] Percentile 61 · #45 · components: 3 provisional
| Source | Component | Value | Elo-eq | Weight | Measured | Variant |
|---|---|---|---|---|---|---|
| LMArena |
vision/overall
|
1,260 ±3.86 | 1,439 ±4 | 63.4% | 2026-10-02 | xhigh |
| LMArena |
vision/ocr
|
1,267 ±4.12 | 1,440 ±4 | 15.6% | 2026-10-02 | xhigh |
| LMArena |
vision/diagram
|
1,275 ±6.16 | 1,441 ±6 | 14.7% | 2026-10-02 | xhigh |
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 1,413 [1,408–1,419] Percentile 71 · #79 · components: 2 provisional
| Source | Component | Value | Elo-eq | Weight | Measured | Variant |
|---|---|---|---|---|---|---|
| LMArena |
text/creative_writing
|
1,395 ±3.96 | 1,423 ±4 | 61.8% | 2026-10-02 | xhigh |
| LMArena |
text/longer_query
|
1,437 ±2.99 | 1,435 ±3 | 32.0% | 2026-10-02 | xhigh |
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: 繁體中文 1,414 [1,406–1,423] Percentile 69 · #69 · components: 1 provisional
| Source | Component | Value | Elo-eq | Weight | Measured | Variant |
|---|---|---|---|---|---|---|
| LMArena |
text/chinese
|
1,469 ±5.84 | 1,429 ±5 | 93.6% | 2026-10-02 | xhigh |
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: Deutsch 1,428 [1,409–1,446] Percentile 78 · #32 · components: 1 provisional
| Source | Component | Value | Elo-eq | Weight | Measured | Variant |
|---|---|---|---|---|---|---|
| LMArena |
text/german
|
1,454 ±12 | 1,449 ±10 | 91.5% | 2026-10-02 | xhigh |
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 1,418 [1,412–1,423] Percentile 72 · #78 · components: 1 provisional
| Source | Component | Value | Elo-eq | Weight | Measured | Variant |
|---|---|---|---|---|---|---|
| LMArena |
text/english
|
1,441 ±2.97 | 1,432 ±3 | 93.9% | 2026-10-02 | xhigh |
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: Español 1,422 [1,406–1,438] Percentile 69 · #47 · components: 1 provisional
| Source | Component | Value | Elo-eq | Weight | Measured | Variant |
|---|---|---|---|---|---|---|
| LMArena |
text/spanish
|
1,446 ±9.52 | 1,441 ±9 | 92.2% | 2026-10-02 | xhigh |
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: Français 1,420 [1,404–1,435] Percentile 61 · #53 · components: 1 provisional
| Source | Component | Value | Elo-eq | Weight | Measured | Variant |
|---|---|---|---|---|---|---|
| LMArena |
text/french
|
1,456 ±9.39 | 1,438 ±9 | 92.3% | 2026-10-02 | xhigh |
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: 日本語 1,417 [1,399–1,436] Percentile 69 · #40 · components: 1 provisional
| Source | Component | Value | Elo-eq | Weight | Measured | Variant |
|---|---|---|---|---|---|---|
| LMArena |
text/japanese
|
1,413 ±14 | 1,437 ±10 | 91.6% | 2026-10-02 | xhigh |
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: 한국어 1,427 [1,412–1,443] Percentile 76 · #34 · components: 1 provisional
| Source | Component | Value | Elo-eq | Weight | Measured | Variant |
|---|---|---|---|---|---|---|
| LMArena |
text/korean
|
1,414 ±11 | 1,446 ±9 | 92.3% | 2026-10-02 | xhigh |
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: Polski 1,393 [1,373–1,412] Percentile 41 · #78 · components: 1 provisional
| Source | Component | Value | Elo-eq | Weight | Measured | Variant |
|---|---|---|---|---|---|---|
| LMArena |
text/polish
|
1,404 ±13 | 1,411 ±11 | 91.3% | 2026-10-02 | xhigh |
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 1,418 [1,410–1,426] Percentile 71 · #68 · components: 1 provisional
| Source | Component | Value | Elo-eq | Weight | Measured | Variant |
|---|---|---|---|---|---|---|
| LMArena |
text/russian
|
1,428 ±4.81 | 1,432 ±4 | 93.7% | 2026-10-02 | xhigh |
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
The same in a table
| Model | 2026-10 |
|---|---|
| GPT-5.6 Luna | 1,438 A |
Badge
Show the model's fedi-index on your site or in a README. The badge updates itself; it links to this page.
Data CC BY 4.0: keep the link to fedi.software.