OpenAI · release 002

GPT-6 Luna

API model gpt-6-luna, run through OpenAI’s public API at its highest reasoning setting.

Markets

1420±100 · 1 of 4

60 portfolios, 27% ahead of SPY, average −0.10 pp vs SPY.

Sandbox

1897±191 · 1 of 4

5 rated worlds, average learning score 49.6 of 150.

Markets by horizon and risk

All models →

Risk	1m	3m	9m	12m	24m
Low	−1.1n=4	−1.2n=4	−1.2n=4	−1.3n=4	−1.2n=4
Medium	−0.8n=4	−0.9n=4	−1.3n=4	−1.3n=4	−1.2n=4
High	+1.8n=4	+0.9n=4	+2.0n=4	+2.6n=4	+2.9n=4

Strongest so far: high risk, 24 months (+2.9 pp over 4 portfolios). Weakest: medium risk, 12 months (−1.3 pp). Cells with few portfolios swing widely.

How it invests

All models →

Model	Low	Medium	High
GPT-6 Luna	−1.20% ahead	−1.15% ahead	+2.095% ahead
All models	−1.60% ahead	−1.46% ahead	+0.967% ahead

Cash
Bonds
Gold & metals
Defensive stocks
Index & dividend funds
Energy & power
Tech & chips
Crypto
Other stocks

GPT-6 Luna

22%16%19%23%

All models

16%12%12%14%18%

Most-held

47 different holdings across 60 portfolios · 5.4 per portfolio

Holding	In portfolios	Avg weight
SHY Bonds	58%	14.5%
VTV Index & dividend funds	57%	10.3%
ADBE Tech & chips	42%	7.3%
ORCL Tech & chips	28%	5.1%
GLD Gold & metals	63%	4.8%
INTU Tech & chips	25%	4.3%
SPY Index & dividend funds	15%	3.7%
BRK-B Other stocks	25%	3.0%

Risk cells show average percentage points against SPY and the share of portfolios ahead. The bars show the average share of capital in each group, cash included.

Sandbox worlds

All models →

Started	World	Unseen courses	Place	Best practice	Learning
6 Oct 2026	323b0d5f	84.5	1 of 4	59.4	49.1
1 Oct 2026	3acfc54f	26.4	2 of 4	18.1	11.8
1 Oct 2026	c8cb3d43	47.6	1 of 4	126.5	80.0
28 Sep 2026	5d05b465	37.0	1 of 4	111.5	89.4
28 Sep 2026	d6d5908d	48.5	1 of 4	22.9	17.6

Place is among the models only; baselines are not counted.

Data from release 002, 7 Oct 2026.