Official run

GPT-6 LunaHigh

In this Twenty Questions LLM benchmark, GPT-6 Luna (high) ranks 14 of 18 with a 21.17 question score and 80% success. Lower is better. The run contains 30 scored episodes across 10 subjects with 3 trials per subject. Choose a subject to inspect its attempts.

Rank 14 of 18openaiHigh

Question score21.17questions · lower is better95% CI 18.18–24.15

Success: 80%
Contract compliance: 100%
Guesser cost: $0.09
Guesser time: 35m 28s

Drill down

Subjects.

01Albert Einstein10.00 avg 02Albert Schweitzer30.67 avg 03Garfield14.33 avg 04Achilles10.00 avg 05Genghis Khan15.00 avg 06Bike pump41.00 avg 07Spider web17.67 avg 08Eyebrow31.33 avg 09Moon9.33 avg 10Door handle32.33 avg

Complete benchmark

Run ledger.

All Guesser and benchmark-support activity.

Total tokens: 5,077,343
Wall-clock runtime: 1h 54m
Guesser calls: 659
Completed: Sep 26, 2026, 8:48:55 PM

Reliability

Output contract clean.

All 659 evaluated outputs matched the public structured-action contract.

Run details

Execution: BX-20260926-B-0003-official-M0028-001
Model ID: M-0028
Benchmark: B-0003
Base seed: 0
Git commit: 6e46fb003ebf

Recorded cost

By role.

Cost composition.Share of full-run cost

Guesser3.5%$0.09
Primary Oracle47.9%$1.26
Reviewer26%$0.68
Judge22.4%$0.59
Validator0.2%$0.01

Roles in this cost:

Guesser: Asks the questions and submits the scored guess.
Primary Oracle: Searches for evidence and proposes an answer.
Reviewer: Checks each Oracle YES or NO independently.
Judge: Decides when the Oracle and Reviewer disagree.
Validator: Checks a submitted guess against the trusted subject.

Read the role and answer-checking method →

Run configuration

Models.

Model identity, prompt versions, routing, calls, and recorded cost for the full run.

GuesserModel under test

openai/gpt-6-luna

Asks the questions and submits the scored guess.

Calls: 659
Cost: $0.0932
Reasoning: High
Routing: Exact provider · openai
Resolved model: openai/gpt-6-luna
Resolved provider: OpenAI
Prompt contract: stateful-category-guesser-v16-five-answer-category-guide
Configuration: M-0028

Resolved provider details

OpenAI659 calls$0.09322,128.14 s

Fallback calls: 0
Provider unreported: 0

Game support

Answering and checking models.

These models support the game. They are fixed across the run and are not under test.

Primary Oracle

openai/gpt-5.6-luna

Searches for evidence and proposes an answer.

Calls: 507
Cost: $1.2593
Reasoning: Medium
Routing: Exact provider · openai
Resolved model: openai/gpt-5.6-luna
Resolved provider: OpenAI
Prompt contract: live-web-oracle-v17-labelled-source-context

Resolved provider details

OpenAI507 calls$1.25933,600.72 s

Fallback calls: 0
Provider unreported: 0

Reviewer

google/gemini-3.5-flash-lite

Checks each Oracle YES or NO independently.

Calls: 494
Cost: $0.6843
Reasoning: Medium
Routing: Exact provider · google-ai-studio
Resolved provider: Google AI Studio

Resolved provider details

Google AI Studio494 calls$0.6843866.40 s

Fallback calls: 0
Provider unreported: 0

Judge

anthropic/claude-opus-5

Decides when the Oracle and Reviewer disagree.

Calls: 28
Cost: $0.5883
Reasoning: Medium
Routing: OpenRouter automatic routing
Resolved provider: Claude Platform on AWS

Resolved provider details

Claude Platform on AWS28 calls$0.5883160.42 s

Fallback calls: 0
Provider unreported: 0

Guess Validator

openai/gpt-5.6-luna

Checks a submitted guess against the trusted subject.

Calls: 33
Cost: $0.0064
Reasoning: Medium
Routing: Exact provider · openai
Resolved model: openai/gpt-5.6-luna
Resolved provider: OpenAI
Prompt contract: strict-guess-validator-v2-generic-kinds
Configuration: gpt-5.6-luna-validator

Resolved provider details

OpenAI33 calls$0.006455.33 s

Fallback calls: 0
Provider unreported: 0