Official run

Gemini 3.8 FlashHigh

In this Twenty Questions LLM benchmark, Gemini 3.8 Flash (high) ranks 2 of 11 with a 14.87 question score and 100% 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 2 of 11google-ai-studioHigh

Question score14.87questions · lower is better95% CI 13.19–16.54

Success: 100%
Contract compliance: 100%
Guesser cost: $2.62
Guesser time: 51m 12s

Drill down

Subjects.

01Albert Einstein11.67 avg 02Albert Schweitzer20.33 avg 03Garfield12.00 avg 04Achilles8.00 avg 05Genghis Khan11.00 avg 06Bike pump27.33 avg 07Spider web11.00 avg 08Eyebrow13.67 avg 09Moon10.67 avg 10Door handle23.00 avg

Complete benchmark

Run ledger.

All Guesser and benchmark-support activity.

Total tokens: 4,646,620
Wall-clock runtime: 2h 58m
Guesser calls: 476
Completed: Sep 8, 2026, 11:14:11 AM

Reliability

Output contract clean.

All 476 evaluated outputs matched the public structured-action contract.

Provenance

Execution: BX-20260908-B-0003-experimental-M0021-001
Model ID: M-0021
Benchmark: B-0003
Base seed: 0
Git commit: eceee31eef2e

Recorded cost

By role.

Cost composition.Share of full-run cost

Guesser58.1%$2.62
Primary Oracle24.1%$1.09
Reviewer12.2%$0.55
Judge5.5%$0.25
Validator0.1%$0.01

Excluded repair overhead$0.032 superseded attempts across 2 affected trials. Not included above.

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

google/gemini-3.8-flash

Asks the questions and submits the scored guess.

Calls: 476
Cost: $2.6240
Reasoning: High
Routing: Exact provider · google-ai-studio
Resolved model: google/gemini-3.8-flash
Resolved provider: Google AI Studio
Prompt contract: stateful-category-guesser-v16-five-answer-category-guide
Configuration: M-0021

Resolved provider details

Google AI Studio476 calls$2.62403,072.10 s

Fallback calls: 0
Provider unreported: 0

Game support

Oracle and adjudication.

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: 442
Cost: $1.0864
Reasoning: Medium
Routing: Exact provider · openai
Resolved model: openai/gpt-5.6-luna
Resolved provider: OpenAI
Prompt contracts: live-web-oracle-v16-search-headroom
live-web-oracle-v17-labelled-source-context

Resolved provider details

OpenAI442 calls$1.08643,743.13 s

Fallback calls: 0
Provider unreported: 0

Reviewer

google/gemini-3.5-flash-lite

Checks each Oracle YES or NO independently.

Calls: 437
Cost: $0.5520
Reasoning: Medium
Routing: Exact provider · google-ai-studio
Resolved provider: Google AI Studio

Resolved provider details

Google AI Studio437 calls$0.5520693.28 s

Fallback calls: 0
Provider unreported: 0

Judge

anthropic/claude-opus-5

Decides when the Oracle and Reviewer disagree.

Calls: 12
Cost: $0.2462
Reasoning: Medium
Routing: OpenRouter automatic routing
Resolved provider: Claude Platform on AWS

Resolved provider details

Claude Platform on AWS12 calls$0.246275.76 s

Fallback calls: 0
Provider unreported: 0

Guess Validator

openai/gpt-5.6-luna

Checks a submitted guess against the trusted subject.

Calls: 34
Cost: $0.0065
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

OpenAI34 calls$0.006556.22 s

Fallback calls: 0
Provider unreported: 0