chat model, answer-letter probabilities · League 2

GLM-5.3low

FinishedPublished

Results f24d1f41, updated 16 November 2026, 21:24 UTC; scored by GNU Backgammon 1.08, 2-ply.

How we called it

Model ID
z-ai/glm-5.3
Provider or endpoint
Fireworks via OpenRouter, quantization not published
Date run
2026-10-19 to 2026-10-26
Price
$1.4 / $4.4 per million tokens in / out; API price; price date not recorded
Model type
chat model, answer-letter probabilities
Weights
open
Runs on
API
Reasoning setting
reasoning effort low (reasoning is mandatory on this endpoint, so no off setup)
Confidence source
letter probability
Rules text
long
Option groups
up to 26 options per request (letters A-Z); more legal plays are split into equal groups, a winner per group, then among the winners; in this data 203 decisions were split into groups of 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26
Sampling settings
temperature 1.0
Notes
OpenRouter, pinned to Fireworks (no fallbacks); quantization not published
Earlier name
GLM-5.3 (on) (renamed on 28 Oct 2026 to the reasoning effort its requests sent)

Missing B7 fields: price.date

Results in League 2

Error rate (PR)
30.927.8 to 34.0
Decisions
1,989 (1,989 scored, 0 invalid)
Engine's best play
50.2%
Blunder rate
21.7%
P95 equity loss
0.310
Mean loss, worst 5%
0.512
Median / P90 s per decision
1.75 / 7.48
$ per 1,000 decisions
2.917API price
Risk judgement, margin 0.001
61.7 (95% interval 59.7 to 63.8); risks found 30.7% (205 / 668), right safety 92.7% (726 / 783)
Risk judgement, margin 0.02
61.6 (95% interval 59.5 to 63.9); risks found 31.8% (184 / 579), right safety 91.4% (797 / 872)
No trade-off
538 (left out of every risk metric)

JSON

Risk call card

61.7risk judgement, margin 0.001

right risks205
wrong risks41
needless risks57
missed risks422
right safety726

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

Close choices

A decision is a close choice when the second most likely play has at least half the probability of the first (internally: split).

Probabilities over plays, not sampled reasoning. Method

League 2: complete letter-readout records are mostly few-option positions, a selected subset.

Decisions
1,989
Eligible (full probability distribution)
527
Left out, by reason
asked in groups of options 184; incomplete distribution (an option without its own probability) 1,090; not audited 187; one letter only, or a program 1
Close choices
5.5% (29 / 527)
Ties in the top two
17 decisions (17 at a positive probability, 0 at 0; the two most likely plays tied: 4, the second tied with the third: 13); 4 of them close choices. Ties are broken by the order the plays were shown, never by the engine.
Between a safe play and a riskier one
31.0% (9 / 29), under 200 decisions of the close choices; 1.7% (9 / 527) of all eligible decisions
Probabilities of the two most likely plays, mean with median
DecisionsNumberp1 (most likely play)p2 (second)p1 + p2
Close choices290.528median 0.5250.382median 0.3920.910median 0.963
Other decisions4980.954median 0.9820.036median 0.0120.990median 0.998
How much probability the two plays hold together: decisions per bin
p1 + p20.0–0.10.1–0.20.2–0.30.3–0.40.4–0.50.5–0.60.6–0.70.7–0.80.8–0.90.9–1.0
Close choices00000034220
Other decisions000000027489
What differs between the two most likely plays (several can differ; none is picked as the cause)
DecisionsNumberHitting rollsNewly made pointsPip countsLongest primeNone / 1 / 2 / 3 / 4 differ
Close choices2931.0%9 / 29, under 200 decisions17.2%5 / 29, under 200 decisions3.4%1 / 29, under 200 decisions17.2%5 / 29, under 200 decisions17 / 7 / 2 / 3 / 0
Other decisions49840.6%202 / 49814.5%72 / 49813.3%66 / 49825.7%128 / 498160 / 223 / 100 / 15 / 0
Engine values (GNUbg 2-ply, equity), mean with median
DecisionsNumberGap between the two plays' valuesMost likely play: behind the engine's bestSecond most likely play: behind the engine's best
Close choices290.113median 0.0410.078median 0.0080.037median 0.000
Other decisions4980.134median 0.0540.026median 0.0000.127median 0.042
Close choices between a safe play and a riskier one, by position class
MarginIn Risk pays positionsIn Safety pays positions
margin 0.001100.0%3 / 3, under 200 decisions60.0%6 / 10, under 200 decisions
margin 0.02100.0%3 / 3, under 200 decisions60.0%6 / 10, under 200 decisions
Close choices among eligible decisions, by position class and stake band
MarginRisk paysSafety paysNo trade-offRisk pays, small stakeRisk pays, medium stakeRisk pays, big stake
margin 0.0014.0%3 / 75, under 200 decisions5.6%10 / 178, under 200 decisions5.8%16 / 2740.0%0 / 12, under 200 decisions8.0%2 / 25, under 200 decisions2.6%1 / 38, under 200 decisions
margin 0.024.8%3 / 63, under 200 decisions5.3%10 / 190, under 200 decisions5.8%16 / 274N/A0 / 08.0%2 / 25, under 200 decisions2.6%1 / 38, under 200 decisions

At the 0.02 margin a Risk pays position has a stake of at least 0.02, so the small band is empty. A big stake means playing safe costs more by the engine's value, not that the decision was harder.

Missed risks where the risk is worth taking, close choices vs other decisions
MarginClose choices: missed risksOther decisions: missed risks (weighted to the close choices' mix)Difference, close minus otherCrude difference
margin 0.00162 decisions0.0%crude 0.0% (0 / 3), under 200 decisions84.7%crude 83.1% (49 / 59), under 200 decisions−84.7 points95% interval −100.0 to −70.0−83.1 points95% interval −90.9 to −74.5
margin 0.0251 decisions0.0%crude 0.0% (0 / 3), under 200 decisions84.7%crude 81.2% (39 / 48), under 200 decisions−84.7 points95% interval −100.0 to −70.0−81.2 points95% interval −90.9 to −71.4

By game phase

PhaseError rate (PR)Decisions
Opening29.0297
Contact36.41,257
Holding gameToo few decisions50
Race21.3254
Bear-offToo few decisions131