chat model, one letter · v1 set (earlier version)

Gemini 4 Argonshuffled

Earlier version

Results f24d1f41, updated 16 November 2026, 22:20 UTC; scored by GNU Backgammon 1.08, 2-ply.

How we called it

Model ID
gemini-4-argon
Provider or endpoint
Gemini API
Date run
Missing
Price
$2.0 / $10.0 per million tokens in / out; API price; price date not recorded
Model type
chat model, one letter
Weights
closed
Runs on
API
Reasoning setting
thinking level low
Confidence source
none
Rules text
short (the v1 rules text; the same stored request text for every model)
Option groups
not recorded; no decision in this data was split
Sampling settings
temperature 1.0
Notes
the same setup with the options in a seeded random order per position, to measure the effect of the v1 play order
Play order
shuffled (replay of the v1 set with the plays in a seeded random order)

Missing B7 fields: option_groups, price.date, date_run

Results in v1 set (earlier version)

Earlier version: plays were listed in the runner's order, which tends to put the best play early. When Gemini 4 Argon answered the same 1,000 positions with the plays shuffled, its error rate changed by +1.9 (95% interval −0.2 to +4.2), so the effect is not proven.

Error rate (PR)
18.916.3 to 21.7
Decisions
1,000 (1,000 scored, 0 invalid)
Engine's best play
61.6%
Blunder rate
15.0%
P95 equity loss
0.204
Mean loss, worst 5%
0.348
Median / P90 s per decision
3.66 / 31.3
$ per 1,000 decisions
Not measured
Risk judgement, margin 0.001
69.2 (95% interval 66.2 to 72.0); risks found 47.3% (168 / 355), right safety 91.1% (336 / 369)
Risk judgement, margin 0.02
68.6 (95% interval 65.3 to 71.7); risks found 49.1% (156 / 318), right safety 88.2% (358 / 406)
No trade-off
276 (left out of every risk metric)

JSON

Risk call card

69.2risk judgement, margin 0.001

right risks168
wrong risks29
needless risks33
missed risks158
right safety336

Hit risk only: being hit. Gammon risk and priming are not in it.

By game phase

PhaseError rate (PR)Decisions
OpeningToo few decisions143
Contact25.7641
Holding gameToo few decisions20
RaceToo few decisions136
Bear-offToo few decisions60