Official run

Claude Fable 5.1High

In this Twenty Questions LLM benchmark, Claude Fable 5.1 (high) ranks 3 of 11 with a 15.10 question score and 93% 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 11anthropicHigh

Question score15.10questions · lower is better95% CI 13.76–16.44

Success: 93%
Contract compliance: 99%
Guesser cost: $8.31
Guesser time: 1h 19m

Drill down

Subjects.

01Albert Einstein9.67 avg 02Albert Schweitzer18.00 avg 03Garfield8.67 avg 04Achilles8.33 avg 05Genghis Khan9.33 avg 06Bike pump38.33 avg 07Spider web16.67 avg 08Eyebrow12.00 avg 09Moon8.67 avg 10Door handle21.33 avg

Complete benchmark

Run ledger.

All Guesser and benchmark-support activity.

Total tokens: 4,602,467
Wall-clock runtime: 3h 8m
Guesser calls: 481
Completed: Sep 8, 2026, 8:37:41 PM

Reliability

Output contract breached.

3 invalid outputs affected 3 episodes and consumed 3 counted turns.

Provenance

Execution: BX-20260908-B-0003-experimental-M0020-001
Model ID: M-0020
Benchmark: B-0003
Base seed: 0
Git commit: eceee31eef2e

Recorded cost

By role.

Cost composition.Share of full-run cost

Guesser81.3%$8.31
Primary Oracle10.1%$1.03
Reviewer5%$0.51
Judge3.6%$0.36
Validator0.1%$0.01

Excluded repair overhead$1.542 superseded attempts across 2 affected trials.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

anthropic/claude-fable-5.1

Asks the questions and submits the scored guess.

Calls: 481
Cost: $8.3148
Reasoning: High
Routing: Exact provider · anthropic
Resolved model: anthropic/claude-fable-5.1
Resolved provider: Anthropic
Prompt contract: stateful-category-guesser-v16-five-answer-category-guide
Configuration: M-0020

Resolved provider details

Anthropic481 calls$8.31484,766.93 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: 398
Cost: $1.0293
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

OpenAI398 calls$1.02933,788.93 s

Fallback calls: 0
Provider unreported: 0

Reviewer

google/gemini-3.5-flash-lite

Checks each Oracle YES or NO independently.

Calls: 388
Cost: $0.5084
Reasoning: Medium
Routing: Exact provider · google-ai-studio
Resolved provider: Google AI Studio

Resolved provider details

Google AI Studio388 calls$0.5084641.25 s

Fallback calls: 0
Provider unreported: 0

Judge

anthropic/claude-opus-5

Decides when the Oracle and Reviewer disagree.

Calls: 18
Cost: $0.3640
Reasoning: Medium
Routing: OpenRouter automatic routing
Resolved provider: Claude Platform on AWS

Resolved provider details

Claude Platform on AWS18 calls$0.3640113.38 s

Fallback calls: 0
Provider unreported: 0

Guess Validator

openai/gpt-5.6-luna

Checks a submitted guess against the trusted subject.

Calls: 43
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

OpenAI43 calls$0.008390.14 s

Fallback calls: 0
Provider unreported: 0