chat model, one letter · League 3
Grok 4.7high
FinishedPublished
Results f24d1f41, updated 16 November 2026, 21:24 UTC; scored by GNU Backgammon 1.08, 2-ply.
How we called it
- Model ID
- grok-4.7
- Provider or endpoint
- xAI subscription (SuperGrok)
- Date run
- 2026-10-20 to 2026-10-24
- Price
- $2.0 / $6.0 per million tokens in / out; API-equivalent (subscription); price date not recorded
- Model type
- chat model, one letter
- Weights
- closed
- Runs on
- subscription
- Reasoning setting
- reasoning effort high, grok-4.7's default
- Confidence source
- none
- 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 195 decisions were split into groups of 14, 15, 16, 17, 18, 19, 20, 21, 22, 23, 24, 25, 26
- Sampling settings
- not sent (API default)
- Notes
- letter constrained by a JSON schema
- Earlier name
- Grok 4.7 (on) (renamed on 28 Oct 2026 to the reasoning effort its requests sent)
Missing B7 fields: price.date
Results in League 3
- Error rate (PR)
- 18.815.9 to 21.8
- Decisions
- 1,337 (1,337 scored, 0 invalid)
- Engine's best play
- 62.5%
- Blunder rate
- 14.0%
- P95 equity loss
- 0.215
- Mean loss, worst 5%
- 0.367
- Median / P90 s per decision
- 122.3 / 588.6
- $ per 1,000 decisions
- 117.379API-equivalent (subscription)
- Risk judgement, margin 0.001
- 68.5 (95% interval 66.1 to 71.1); risks found 44.0% (190 / 432), right safety 93.0% (454 / 488)
- Risk judgement, margin 0.02
- 67.9 (95% interval 64.9 to 70.9); risks found 44.6% (171 / 383), right safety 91.1% (489 / 537)
- No trade-off
- 417 (left out of every risk metric)
Risk call card
68.5risk judgement, margin 0.001
right risks190
wrong risks24
needless risks34
missed risks218
right safety454
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
- Decisions
- 1,337
- Eligible (full probability distribution)
- 0
- Left out, by reason
- one letter only, or a program 1,337
- Close choices
- N/A no full probability distribution
By game phase
| Phase | Error rate (PR) | Decisions |
|---|---|---|
| Opening | 17.6 | 201 |
| Contact | 27.5 | 724 |
| Holding game | Too few decisions | 101 |
| Race | Too few decisions | 161 |
| Bear-off | Too few decisions | 150 |