Official run

Claude Opus 5.5High

In this Twenty Questions LLM benchmark, Claude Opus 5.5 (high) ranks 3 of 18 with a 15.00 question score and 97% 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 3 of 18anthropicHigh

Question score15.00questions · lower is better95% CI 13.24–16.76

Success: 97%
Contract compliance: 100%
Guesser cost: $2.25
Guesser time: 38m 35s

Drill down

Subjects.

01Albert Einstein9.00 avg 02Albert Schweitzer20.67 avg 03Garfield13.33 avg 04Achilles7.00 avg 05Genghis Khan9.33 avg 06Bike pump32.67 avg 07Spider web18.33 avg 08Eyebrow9.33 avg 09Moon7.00 avg 10Door handle23.33 avg

Complete benchmark

Run ledger.

All Guesser and benchmark-support activity.

Total tokens: 3,663,915
Wall-clock runtime: 3h 51m
Guesser calls: 479
Completed: Sep 26, 2026, 11:33:57 PM

Reliability

Output contract clean.

All 479 evaluated outputs matched the public structured-action contract.

Run details

Execution: BX-20260926-B-0003-official-M0030-001
Model ID: M-0030
Benchmark: B-0003
Base seed: 0
Git commit: 6e46fb003ebf

Recorded cost

By role.

Cost composition.Share of full-run cost

Guesser58.6%$2.25
Primary Oracle20.8%$0.80
Reviewer10.7%$0.41
Judge9.8%$0.37
Validator0.2%$0.01

Excluded repair overhead$0.111 superseded attempt across 1 affected trial. 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

anthropic/claude-opus-5.5

Asks the questions and submits the scored guess.

Calls: 479
Cost: $2.2516
Reasoning: High
Routing: Exact provider · anthropic
Resolved model: anthropic/claude-opus-5.5
Resolved provider: Anthropic
Prompt contract: stateful-category-guesser-v16-five-answer-category-guide
Configuration: M-0030

Resolved provider details

Anthropic479 calls$2.25162,314.96 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: 325
Cost: $0.7984
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

OpenAI325 calls$0.79842,354.94 s

Fallback calls: 0
Provider unreported: 0

Reviewer

google/gemini-3.5-flash-lite

Checks each Oracle YES or NO independently.

Calls: 307
Cost: $0.4103
Reasoning: Medium
Routing: Exact provider · google-ai-studio
Resolved provider: Google AI Studio

Resolved provider details

Google AI Studio307 calls$0.4103510.68 s

Fallback calls: 0
Provider unreported: 0

Judge

anthropic/claude-opus-5

Decides when the Oracle and Reviewer disagree.

Calls: 18
Cost: $0.3748
Reasoning: Medium
Routing: OpenRouter automatic routing
Resolved provider: Claude Platform on AWS

Resolved provider details

Claude Platform on AWS18 calls$0.3748101.51 s

Fallback calls: 0
Provider unreported: 0

Guess Validator

openai/gpt-5.6-luna

Checks a submitted guess against the trusted subject.

Calls: 42
Cost: $0.0083
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

OpenAI42 calls$0.008368.78 s

Fallback calls: 0
Provider unreported: 0