Head-to-head comparison

Claude Fable 5.1 vs. GPT-6 Astra

Compare published benchmark results from matching versions and cohorts, API pricing, and technical specifications. Any documented test setup differences remain visible.

shared result series: 228
distinct benchmark metrics: 40
latest retrieval: 2026-10-15

First modelSecond model

Specifications

General specifications

Release date, availability, context window, architecture, and published parameter counts at a glance. Undisclosed details remain marked as unavailable.

Specifications for Claude Fable 5.1 and GPT-6 Astra
Attribute	Claude Fable 5.1	GPT-6 Astra
Provider	Anthropic	OpenAI
Released	Sep 1, 2026	Sep 3, 2026
Availability	Active	Limited access
Model type	Proprietary	Proprietary
Parameters	Not published	Not published
Architecture	Not published	Not published
Context window	1,000,000 tokens	1,050,000 tokens
Knowledge cutoff	June 2026	April 30, 2026
Technical sources	Model	Model, Context

Pricing

API pricing compared

The table compares official standard API rates per 1M tokens, including published cache pricing. Batch, fast, priority, regional surcharges, tool calls, and cloud platform pricing are excluded.

API pricing for Claude Fable 5.1 and GPT-6 Astra
Price type	Claude Fable 5.1	GPT-6 Astra
API input	$10	$10 ≤272K / $20 >272K
API output	$50	$50 ≤272K / $75 >272K
Cache read	$0.25	$1 ≤272K / $2 >272K
Cache write (5 min.)	$12.5	$12.5 ≤272K / $25 >272K
Cache write (1 hr.)	$20	Not listed
Price verified	09/22/2026 Source	09/22/2026 Source

Claude Fable 5.1GPT-6 Astra

API inputUSD per 1M tokens, base tier

Claude Fable 5.1$10

GPT-6 Astra$10

API outputUSD per 1M tokens, base tier

Claude Fable 5.1$50

GPT-6 Astra$50

Cache readUSD per 1M tokens, base tier

Claude Fable 5.1$0.25

GPT-6 Astra$1

Key to the marks:better valuebetter, but not conclusivetie

Pricing by provider

Direct developer pricing and provider endpoint tariffs routed through OpenRouter are listed separately. Prices are in USD per 1M tokens. Context-dependent direct tiers plus regional, Flex, Priority, and other OpenRouter tariffs remain individually identifiable.

Provider pricing for Claude Fable 5.1 and GPT-6 Astra
Model	Provider	Purchase route and tariff	Input per 1M tokens	Output per 1M tokens	Source
Claude Fable 5.1	Anthropic	Direct from developerStandard API	Lowest listed price: $10	Lowest listed price: $50	Source09/22/2026
	Amazon Bedrock	Via OpenRouteramazon-bedrock	Lowest listed price: $10	Lowest listed price: $50	Source10/19/2026
	Anthropic	Via OpenRouteranthropic	Lowest listed price: $10	Lowest listed price: $50	Source10/19/2026
	Azure	Via OpenRouterazure	Lowest listed price: $10	Lowest listed price: $50	Source10/19/2026
	Google	Via OpenRoutergoogle-vertex/global	Lowest listed price: $10	Lowest listed price: $50	Source10/19/2026
GPT-6 Astra	OpenAI	Direct from developerUp to 272,000 context tokens	$10	$50	Source09/22/2026
	OpenAI	Direct from developerAbove 272,000 context tokens	$20	$75	Source09/22/2026
	Amazon Bedrock	Via OpenRouteramazon-bedrock/us-west-2	$11	$55	Source10/19/2026
	Azure	Via OpenRouterazure	$10	$50	Source10/19/2026
	Azure	Via OpenRouterazure/us	$11	$55	Source10/19/2026
	OpenAI	Via OpenRouteropenai	$10	$50	Source10/19/2026
	OpenAI	Via OpenRouteropenai/fast	$20	$100	Source10/19/2026
	OpenAI	Via OpenRouteropenai/flex	Lowest listed price: $5	Lowest listed price: $25	Source10/19/2026
	OpenAI	Via OpenRouteropenai/ultrafast	$60	$300	Source10/19/2026

Performance

Shared benchmarks

Only published results with the same benchmark version, task, metric, and comparison cohort are paired. Different reasoning levels, agents, harnesses, or output limits appear directly in the table.

Capability profile from matched benchmarks

Each axis averages directly matched benchmark families. A value of 100 means the stronger value within this pair, not a universal quality score.

Claude Fable 5.1

GPT-6 Astra

Coding: 96 / 98 (6 benchmark families)

Reasoning: 100 / 99.1 (1 benchmark family)

Mathematics: 100 / 97.6 (1 benchmark family)

Knowledge: 98.5 / 99.3 (2 benchmark families)

Finance: 100 / 95 (2 benchmark families)

Legal: 87.2 / 99.8 (2 benchmark families)

Agentic tasks: 92.4 / 91.2 (3 benchmark families)

Multimodal: 99.6 / 99.8 (1 benchmark family)

55 matched rows from 17 benchmark families were considered. The radar shows 8 of 12 comparable categories.

Claude Fable 5.1GPT-6 Astra

AutomationBench-AA (strict pass@1)Overall. Higher is better. No documented setup difference

Claude Fable 5.132.1 %

GPT-6 Astra41.6 %

Code MigrationOverall. Higher is better. Test setup not documented as identical

Claude Fable 5.154.607 %

GPT-6 Astra67.74 %

CyberBench v1.1Overall. Higher is better. Test setup not documented as identical

Claude Fable 5.170.416 %

GPT-6 Astra41.072 %

DeepSWEOverall. Higher is better. No documented setup difference

Claude Fable 5.167.4 %

GPT-6 Astra74.1 %

EMBOverall. Higher is better. Test setup not documented as identical

Claude Fable 5.176.67 %

GPT-6 Astra71.7 %

Finance Agent (v2)Overall. Higher is better. Test setup not documented as identical

Claude Fable 5.158.877 %

GPT-6 Astra53.54 %

GDP.pdfOverall. Higher is better. No documented setup difference

Claude Fable 5.127.6 points

GPT-6 Astra34.2 points

Harvey LAB-AAOverall. Higher is better. No documented setup difference

Claude Fable 5.16.4 %

GPT-6 Astra8.6 %

Harvey's Legal Agent BenchmarkOverall · Task fully resolved. Higher is better. Test setup not documented as identical

Claude Fable 5.16.667 %

GPT-6 Astra5.417 %

HealthBench ProOverall. Higher is better. No documented setup difference

Claude Fable 5.162.1 points

GPT-6 Astra65.3 points

IOIOverall. Higher is better. Test setup not documented as identical

Claude Fable 5.190.778 %

GPT-6 Astra100 %

JobBenchOverall. Higher is better. No documented setup difference

Claude Fable 5.173.1 points

GPT-6 Astra66.1 points

Key to the marks:better valuebetter, but not conclusivetie

Benchmark scores for Claude Fable 5.1 and GPT-6 Astra
Benchmark and task	Claude Fable 5.1	GPT-6 Astra	Source
AutomationBench-AA (strict pass@1)Overall Test details No documented difference in the test setupVersion: Meta Wonder scorecard; version unspecifiedMetric: strict pass@1; every end-state assertion passedScoring: Higher is betterSettings: Reasoning: maxBusiness workflows in a simulated application environment. Every end-state assertion must pass; max reasoning. Do not merge with other AutomationBench-AA metrics.	32.1 % Comparison table published by Meta. Depending on the benchmark, Meta uses internal runs, provider reports or third-party results and selects the highest available validated score. This is not a uniformly independently measured cohort.	Winner41.6 % Comparison table published by Meta. Depending on the benchmark, Meta uses internal runs, provider reports or third-party results and selects the highest available validated score. This is not a uniformly independently measured cohort.	Meta Source details Data as of: 10/15/2026
Code MigrationOverall Test details Test setup not documented as identicalNo direct winnerVersion: 1Metric: accuracyScoring: Higher is betterClaude Fable 5.1: Compute: max, Temperature: 1, Output limit: 128,000 tokens GPT-6 Astra: Reasoning: max, Output limit: 128,000 tokens	54.607 %	Higher measured value67.74 %	Vals AI Source details Data as of: 10/12/2026
CyberBench v1.1Overall Test details Test setup not documented as identicalNo direct winnerVersion: 1.1Metric: accuracyScoring: Higher is betterClaude Fable 5.1: Compute: max, Temperature: 1, Output limit: 128,000 tokens GPT-6 Astra: Reasoning: max, Output limit: 128,000 tokens	Higher measured value70.416 %	41.072 %	Vals AI Source details Data as of: 10/12/2026
DeepSWEOverall Test details No documented difference in the test setupVersion: v1.1Metric: pass@1Scoring: Higher is betterSettings: Reasoning: max113 tasks across 91 repositories and five languages. Max reasoning. Meta attributes these results to Datacurve; the Muse scores could not be found on its public leaderboard at retrieval.	67.4 % Comparison table published by Meta. Depending on the benchmark, Meta uses internal runs, provider reports or third-party results and selects the highest available validated score. This is not a uniformly independently measured cohort.	Winner74.1 % Comparison table published by Meta. Depending on the benchmark, Meta uses internal runs, provider reports or third-party results and selects the highest available validated score. This is not a uniformly independently measured cohort.	Meta Source details Data as of: 10/15/2026
EMBOverall Test details Test setup not documented as identicalNo direct winnerVersion: 1Metric: accuracyScoring: Higher is betterClaude Fable 5.1: Compute: max, Temperature: 1, Output limit: 128,000 tokens GPT-6 Astra: Reasoning: max, Output limit: 128,000 tokens	Higher measured value76.67 %	71.7 %	Vals AI Source details Data as of: 10/12/2026
Finance Agent (v2)Overall Test details Test setup not documented as identicalNo direct winnerVersion: 2Metric: accuracyScoring: Higher is betterClaude Fable 5.1: Compute: max, Temperature: 1, Output limit: 128,000 tokens GPT-6 Astra: Reasoning: max, Output limit: 128,000 tokens	Higher measured value58.877 %	53.54 %	Vals AI Source details Data as of: 10/12/2026
GDP.pdfOverall Test details No documented difference in the test setupVersion: Meta Wonder scorecard; version unspecifiedMetric: published GDP.pdf scoreScoring: Higher is betterSettings: Reasoning: maxProfessional work grounded in supplied PDFs. Max reasoning. Meta names Surge as the original source.	27.6 points Comparison table published by Meta. Depending on the benchmark, Meta uses internal runs, provider reports or third-party results and selects the highest available validated score. This is not a uniformly independently measured cohort. The source specifies neither a denominator nor a normalized scale. This is the published raw score, not a percentage.	Winner34.2 points Comparison table published by Meta. Depending on the benchmark, Meta uses internal runs, provider reports or third-party results and selects the highest available validated score. This is not a uniformly independently measured cohort. The source specifies neither a denominator nor a normalized scale. This is the published raw score, not a percentage.	Meta Source details Data as of: 10/15/2026
Harvey LAB-AAOverall Test details No documented difference in the test setupVersion: 1.1Metric: all criteria passed with hallucination checksScoring: Higher is betterSettings: Reasoning: maxLegal deliverables must pass every criterion including hallucination checks. Max reasoning. Meta attributes the original evaluation to Artificial Analysis.	6.4 % Comparison table published by Meta. Depending on the benchmark, Meta uses internal runs, provider reports or third-party results and selects the highest available validated score. This is not a uniformly independently measured cohort.	Winner8.6 % Comparison table published by Meta. Depending on the benchmark, Meta uses internal runs, provider reports or third-party results and selects the highest available validated score. This is not a uniformly independently measured cohort.	Meta Source details Data as of: 10/15/2026
Harvey's Legal Agent BenchmarkOverall · Task fully resolved Test details Test setup not documented as identicalNo direct winnerVersion: 1Metric: task resolution rateScoring: Higher is betterClaude Fable 5.1: Compute: max, Temperature: 1, Output limit: 128,000 tokens GPT-6 Astra: Reasoning: max, Output limit: 128,000 tokens	Higher measured value6.667 %	5.417 %	Vals AI Source details Data as of: 10/12/2026
HealthBench ProOverall Test details No documented difference in the test setupVersion: Meta Wonder scorecard; version unspecifiedMetric: length-normalized weighted rubric scoreScoring: Higher is betterSettings: Reasoning: max525 health questions without tools; physician rubrics graded by GPT-5.4. Max reasoning. Meta attributes the Claude scores to Anthropic system cards.	62.1 points Comparison table published by Meta. Depending on the benchmark, Meta uses internal runs, provider reports or third-party results and selects the highest available validated score. This is not a uniformly independently measured cohort. The source specifies neither a denominator nor a normalized scale. This is the published raw score, not a percentage.	Winner65.3 points Comparison table published by Meta. Depending on the benchmark, Meta uses internal runs, provider reports or third-party results and selects the highest available validated score. This is not a uniformly independently measured cohort. The source specifies neither a denominator nor a normalized scale. This is the published raw score, not a percentage.	Meta Source details Data as of: 10/15/2026
IOIOverall Test details Test setup not documented as identicalNo direct winnerVersion: 2Metric: accuracyScoring: Higher is betterClaude Fable 5.1: Compute: max, Temperature: 1, Output limit: 128,000 tokens GPT-6 Astra: Reasoning: max, Output limit: 128,000 tokens	90.778 %	Higher measured value100 %	Vals AI Source details Data as of: 10/14/2026
JobBenchOverall Test details No documented difference in the test setupVersion: Meta Wonder scorecard; version unspecifiedMetric: mean task rewardScoring: Higher is betterSettings: Reasoning: max65 professional tasks across 35 occupations. Mean task-specific reward, max reasoning.	Winner73.1 points Comparison table published by Meta. Depending on the benchmark, Meta uses internal runs, provider reports or third-party results and selects the highest available validated score. This is not a uniformly independently measured cohort. The source specifies neither a denominator nor a normalized scale. This is the published raw score, not a percentage.	66.1 points Comparison table published by Meta. Depending on the benchmark, Meta uses internal runs, provider reports or third-party results and selects the highest available validated score. This is not a uniformly independently measured cohort. The source specifies neither a denominator nor a normalized scale. This is the published raw score, not a percentage.	Meta Source details Data as of: 10/15/2026
Legal Research BenchOverall · All-pass Test details Test setup not documented as identicalNo direct winnerVersion: 1Metric: all-pass rateScoring: Higher is betterClaude Fable 5.1: Compute: max, Temperature: 1, Output limit: 128,000 tokens GPT-6 Astra: Reasoning: max, Output limit: 128,000 tokens	Higher measured value55.288 %	39.423 %	Vals AI Source details Data as of: 10/12/2026
LMArena Code Arena WebDevOverall Test details No documented difference in the test setupVersion: webdevMetric: WebDev Arena Bradley-Terry ratingScoring: Higher is betterClaude Fable 5.1: Snapshot: claude-fable-5.1-max, Reasoning: max, Harness: Arena Code Arena WebDev GPT-6 Astra: Snapshot: gpt-6-astra-max, Reasoning: max, Harness: Arena Code Arena WebDevClaude Fable 5.1: Arena rating from 7,230 published pairwise votes.GPT-6 Astra: Arena rating from 5,746 published pairwise votes.	1,742.652 95% confidence interval: ±9.147The Arena score measures human preference rather than a fixed capability test. The confidence interval, Arena category, and model configuration are part of the result.	Winner1,780.244 95% confidence interval: ±10.541The Arena score measures human preference rather than a fixed capability test. The confidence interval, Arena category, and model configuration are part of the result.	LMArena Source details Data as of: 10/13/2026
LMArena Text, Style ControlOverall Test details No documented difference in the test setupVersion: style-control-v1Metric: style-controlled Bradley-Terry ratingScoring: Higher is betterClaude Fable 5.1: Snapshot: claude-fable-5.1-max, Reasoning: max, Harness: Arena Text, style control GPT-6 Astra: Snapshot: gpt-6-astra-max, Reasoning: max, Harness: Arena Text, style controlClaude Fable 5.1: Arena rating from 14,121 published pairwise votes.GPT-6 Astra: Arena rating from 11,674 published pairwise votes.	Winner1,500.568 95% confidence interval: ±6.106The Arena score measures human preference rather than a fixed capability test. The confidence interval, Arena category, and model configuration are part of the result.	1,476.716 95% confidence interval: ±6.385The Arena score measures human preference rather than a fixed capability test. The confidence interval, Arena category, and model configuration are part of the result.	LMArena Source details Data as of: 10/13/2026
LMArena Vision, Style ControlOverall Test details No documented difference in the test setupNo clear leadVersion: style-control-v1Metric: style-controlled Bradley-Terry ratingScoring: Higher is betterClaude Fable 5.1: Snapshot: claude-fable-5.1-max, Reasoning: max, Harness: Arena Vision, style control GPT-6 Astra: Snapshot: gpt-6-astra-max, Reasoning: max, Harness: Arena Vision, style controlClaude Fable 5.1: Arena rating from 4,648 published pairwise votes.GPT-6 Astra: Arena rating from 3,590 published pairwise votes.	No clear lead1,286.407 95% confidence interval: ±9.66The Arena score measures human preference rather than a fixed capability test. The confidence interval, Arena category, and model configuration are part of the result.	1,285.988 95% confidence interval: ±10.782The Arena score measures human preference rather than a fixed capability test. The confidence interval, Arena category, and model configuration are part of the result.	LMArena Source details Data as of: 10/09/2026
MedCodeOverall Test details Test setup not documented as identicalNo direct winnerVersion: 1Metric: accuracyScoring: Higher is betterClaude Fable 5.1: Compute: max, Temperature: 1, Output limit: 128,000 tokens GPT-6 Astra: Reasoning: max, Output limit: 128,000 tokens	Higher measured value53.509 %	48.486 %	Vals AI Source details Data as of: 10/13/2026
MedScribeOverall Test details Test setup not documented as identicalNo direct winnerVersion: 1Metric: accuracyScoring: Higher is betterClaude Fable 5.1: Compute: max, Temperature: 1, Output limit: 128,000 tokens GPT-6 Astra: Reasoning: max, Output limit: 128,000 tokens	Higher measured value91.294 %	87.908 %	Vals AI Source details Data as of: 10/13/2026
MemoryBenchOverall Test details No documented difference in the test setupVersion: Meta internal Wonder evaluationMetric: LLM-judge weighted rubric scoreScoring: Higher is betterSettings: Reasoning: maxInternal Meta evaluation with 64K-100K-token conversations, distractors and conflicting or updated turns. Max reasoning.	79 points Comparison table published by Meta. Depending on the benchmark, Meta uses internal runs, provider reports or third-party results and selects the highest available validated score. This is not a uniformly independently measured cohort. The source specifies neither a denominator nor a normalized scale. This is the published raw score, not a percentage.	Winner81.4 points Comparison table published by Meta. Depending on the benchmark, Meta uses internal runs, provider reports or third-party results and selects the highest available validated score. This is not a uniformly independently measured cohort. The source specifies neither a denominator nor a normalized scale. This is the published raw score, not a percentage.	Meta Source details Data as of: 10/15/2026
MysteryMechanismOverall Test details Test setup not documented as identicalNo direct winnerVersion: 1Metric: accuracyScoring: Higher is betterClaude Fable 5.1: Compute: max, Temperature: 1, Output limit: 128,000 tokens GPT-6 Astra: Reasoning: max, Output limit: 128,000 tokens	47.748 %	Higher measured value53.153 %	Vals AI Source details Data as of: 10/12/2026
OSWorld V2.1 (binary completion)Overall Test details No documented difference in the test setupVersion: V2.1Metric: strict binary task completionScoring: Higher is betterSettings: Reasoning: max108 Ubuntu desktop workflows. Full task completion without partial credit, max reasoning. No binary Opus 5.5 score is published.	Winner43.1 % Comparison table published by Meta. Depending on the benchmark, Meta uses internal runs, provider reports or third-party results and selects the highest available validated score. This is not a uniformly independently measured cohort.	34.3 % Comparison table published by Meta. Depending on the benchmark, Meta uses internal runs, provider reports or third-party results and selects the highest available validated score. This is not a uniformly independently measured cohort.	Meta Source details Data as of: 10/15/2026
OSWorld V2.1 (partial reward)Overall Test details No documented difference in the test setupVersion: V2.1Metric: mean per-task partial rewardScoring: Higher is betterSettings: Reasoning: max108 Ubuntu desktop workflows. Partial reward, max reasoning. Meta attributes the Opus 5.5 score to Anthropic.	Winner80.7 % Comparison table published by Meta. Depending on the benchmark, Meta uses internal runs, provider reports or third-party results and selects the highest available validated score. This is not a uniformly independently measured cohort.	75.8 % Comparison table published by Meta. Depending on the benchmark, Meta uses internal runs, provider reports or third-party results and selects the highest available validated score. This is not a uniformly independently measured cohort.	Meta Source details Data as of: 10/15/2026
Personal Superintelligence IndexOverall Test details No documented difference in the test setupVersion: Meta Wonder launch scorecardMetric: Meta internal composite scoreScoring: Higher is betterSettings: Reasoning: maxMeta internal composite of seven personal and professional work evaluations. Max reasoning.	49.9 points Comparison table published by Meta. Depending on the benchmark, Meta uses internal runs, provider reports or third-party results and selects the highest available validated score. This is not a uniformly independently measured cohort. The source specifies neither a denominator nor a normalized scale. This is the published raw score, not a percentage.	Winner54.5 points Comparison table published by Meta. Depending on the benchmark, Meta uses internal runs, provider reports or third-party results and selects the highest available validated score. This is not a uniformly independently measured cohort. The source specifies neither a denominator nor a normalized scale. This is the published raw score, not a percentage.	Meta Source details Data as of: 10/15/2026
ProgramBenchOverall Test details Test setup not documented as identicalNo direct winnerVersion: 1Metric: fully resolved rateScoring: Higher is betterClaude Fable 5.1: Compute: max, Temperature: 1, Output limit: 128,000 tokens GPT-6 Astra: Reasoning: max, Output limit: 128,000 tokens	Higher measured value7 %	5.5 %	Vals AI Source details Data as of: 10/12/2026
ProgramBenchOverall Test details No documented difference in the test setupVersion: Meta Wonder scorecard; version unspecifiedMetric: rollouts passing at least 95% of testsScoring: Higher is betterSettings: Reasoning: max, Harness: mini-swe-agentReconstruct programs from a specification. Common mini-swe-agent, clean environment without internet after setup. Share of rollouts passing at least 95% of tests, max reasoning.	Winner55.5 % Comparison table published by Meta. Depending on the benchmark, Meta uses internal runs, provider reports or third-party results and selects the highest available validated score. This is not a uniformly independently measured cohort.	52.5 % Comparison table published by Meta. Depending on the benchmark, Meta uses internal runs, provider reports or third-party results and selects the highest available validated score. This is not a uniformly independently measured cohort.	Meta Source details Data as of: 10/15/2026
ProofBenchOverall Test details Test setup not documented as identicalNo direct winnerVersion: 1.1Metric: accuracyScoring: Higher is betterClaude Fable 5.1: Compute: max, Temperature: 1, Output limit: 128,000 tokens GPT-6 Astra: Reasoning: max, Output limit: 128,000 tokens	Higher measured value100 %	99 %	Vals AI Source details Data as of: 10/12/2026
SAGEOverall Test details Test setup not documented as identicalNo direct winnerVersion: 1Metric: accuracyScoring: Higher is betterClaude Fable 5.1: Compute: max, Temperature: 1, Output limit: 30,000 tokens GPT-6 Astra: Reasoning: max, Output limit: 30,000 tokens	Higher measured value48.526 %	46.367 %	Vals AI Source details Data as of: 10/08/2026
SRE BenchOverall Test details Test setup not documented as identicalNo direct winnerVersion: 1Metric: fully solved rateScoring: Higher is betterClaude Fable 5.1: Compute: max, Temperature: 1, Output limit: 128,000 tokens GPT-6 Astra: Reasoning: max, Output limit: 128,000 tokens	22.901 %	Higher measured value56.87 %	Vals AI Source details Data as of: 10/14/2026
SWE-bench MultimodalOverall Test details Test setup not documented as identicalNo direct winnerVersion: 1Metric: accuracyScoring: Higher is betterClaude Fable 5.1: Compute: max, Temperature: 1, Output limit: 128,000 tokens GPT-6 Astra: Reasoning: max, Output limit: 128,000 tokens	Higher measured value61.25 %	43.333 %	Vals AI Source details Data as of: 10/13/2026
SWE-Marathon v1.1Overall Test details No documented difference in the test setupVersion: v1.1Metric: pass@1Scoring: Higher is betterSettings: Reasoning: maxComparator score in Meta's published table, max reasoning. The agent is not documented for the scorecard cell.	Winner45.6 % Comparison table published by Meta. Depending on the benchmark, Meta uses internal runs, provider reports or third-party results and selects the highest available validated score. This is not a uniformly independently measured cohort.	42.5 % Comparison table published by Meta. Depending on the benchmark, Meta uses internal runs, provider reports or third-party results and selects the highest available validated score. This is not a uniformly independently measured cohort.	Meta Source details Data as of: 10/15/2026
Tax Agent BenchOverall Test details Test setup not documented as identicalNo direct winnerVersion: 1Metric: accuracyScoring: Higher is betterClaude Fable 5.1: Compute: max, Temperature: 1, Output limit: 128,000 tokens GPT-6 Astra: Reasoning: max, Output limit: 128,000 tokens	Higher measured value49.194 %	20.717 %	Vals AI Source details Data as of: 10/12/2026
Terminal-Bench 2.1Overall Test details Test setup not documented as identicalNo direct winnerVersion: 2.1Metric: accuracyScoring: Higher is betterClaude Fable 5.1: Compute: high, Temperature: 1, Output limit: 128,000 tokens GPT-6 Astra: Reasoning: max, Output limit: 128,000 tokens	85.019 %	Higher measured value87.266 %	Vals AI Source details Data as of: 09/28/2026
Terminal-Bench 4.0Overall Test details Test setup not documented as identicalNo direct winnerVersion: 4.0Metric: accuracyScoring: Higher is betterClaude Fable 5.1: Compute: max, Temperature: 1, Output limit: 128,000 tokens GPT-6 Astra: Reasoning: max, Output limit: 128,000 tokens	58.081 %	Higher measured value59.596 %	Vals AI Source details Data as of: 10/13/2026
Terminal-Bench 4.0-AAOverall Test details No documented difference in the test setupVersion: 4.0-AAMetric: task accuracyScoring: Higher is betterSettings: Reasoning: maxTerminal tasks in isolated environments with executable verifiers. Max reasoning. Meta names Artificial Analysis as the source; each agent remains part of the result.	52 % Comparison table published by Meta. Depending on the benchmark, Meta uses internal runs, provider reports or third-party results and selects the highest available validated score. This is not a uniformly independently measured cohort.	Winner59.1 % Comparison table published by Meta. Depending on the benchmark, Meta uses internal runs, provider reports or third-party results and selects the highest available validated score. This is not a uniformly independently measured cohort.	Meta Source details Data as of: 10/15/2026
Terminal-Bench ScienceOverall Test details Test setup not documented as identicalNo direct winnerVersion: 0.1Metric: accuracyScoring: Higher is betterClaude Fable 5.1: Compute: max, Harness: Mini-SWE-agent, Temperature: 1, Output limit: 128,000 tokens GPT-6 Astra: Reasoning: max, Harness: Mini-SWE-agent, Output limit: 128,000 tokens	40 %	Higher measured value62.857 %	Vals AI Source details Data as of: 10/14/2026
Time Horizon Index: KSPOverall Test details Test setup not documented as identicalNo direct winnerVersion: 1Metric: progressScoring: Higher is betterClaude Fable 5.1: Compute: max, Temperature: 1, Output limit: 128,000 tokens GPT-6 Astra: Reasoning: max, Output limit: 128,000 tokens	63.333 %	Higher measured value90.5 %	Vals AI Source details Data as of: 10/07/2026
Vals Finance Agent V2Overall Test details No documented difference in the test setupVersion: V2Metric: private test accuracyScoring: Higher is betterSettings: Reasoning: maxFinancial research using filings and market data. Private test set, max reasoning. Vals results republished by Meta.	Winner58.9 % Comparison table published by Meta. Depending on the benchmark, Meta uses internal runs, provider reports or third-party results and selects the highest available validated score. This is not a uniformly independently measured cohort.	53.5 % Comparison table published by Meta. Depending on the benchmark, Meta uses internal runs, provider reports or third-party results and selects the highest available validated score. This is not a uniformly independently measured cohort.	Meta Source details Data as of: 10/15/2026
Vals IndexOverall Test details Test setup not documented as identicalNo direct winnerVersion: 2.1Metric: weighted index scoreScoring: Higher is betterClaude Fable 5.1: Compute: max, Temperature: 1, Output limit: 128,000 tokens GPT-6 Astra: Reasoning: max, Output limit: 128,000 tokens	Higher measured value65.827 %	63.125 %	Vals AI Source details Data as of: 10/13/2026
Vibe Code Bench 1-100Overall Test details Test setup not documented as identicalNo direct winnerVersion: 1.0Metric: accuracyScoring: Higher is betterClaude Fable 5.1: Compute: max, Harness: OpenHands, Temperature: 1, Output limit: 128,000 tokens GPT-6 Astra: Reasoning: max, Harness: OpenHands, Output limit: 128,000 tokens	Higher measured value28 %	27.644 %	Vals AI Source details Data as of: 10/14/2026
Vibe Code Bench v1.1Overall Test details Test setup not documented as identicalNo direct winnerVersion: 1.1Metric: accuracyScoring: Higher is betterClaude Fable 5.1: Compute: max, Harness: OpenHands, Temperature: 1, Output limit: 128,000 tokens GPT-6 Astra: Reasoning: max, Harness: OpenHands, Output limit: 128,000 tokens	Higher measured value90.263 %	89.594 %	Vals AI Source details Data as of: 10/12/2026
Code MigrationCLI Test details Test setup not documented as identicalNo direct winnerVersion: 1Metric: accuracyScoring: Higher is betterClaude Fable 5.1: Compute: max, Temperature: 1, Output limit: 128,000 tokens GPT-6 Astra: Reasoning: max, Output limit: 128,000 tokens	52.809 %	Higher measured value67.463 %	Vals AI Source details Data as of: 10/12/2026
CyberBench v1.1Patch Test details Test setup not documented as identicalNo direct winnerVersion: 1.1Metric: accuracyScoring: Higher is betterClaude Fable 5.1: Compute: max, Temperature: 1, Output limit: 128,000 tokens GPT-6 Astra: Reasoning: max, Output limit: 128,000 tokens	Higher measured value87.5 %	82.143 %	Vals AI Source details Data as of: 10/12/2026
EMBComps Test details Test setup not documented as identicalNo direct winnerVersion: 1Metric: accuracyScoring: Higher is betterClaude Fable 5.1: Compute: max, Temperature: 1, Output limit: 128,000 tokens GPT-6 Astra: Reasoning: max, Output limit: 128,000 tokens	Higher measured value82.233 %	77.953 %	Vals AI Source details Data as of: 10/12/2026
Finance Agent (v2)Adjustments Test details Test setup not documented as identicalNo direct winnerVersion: 2Metric: accuracyScoring: Higher is betterClaude Fable 5.1: Compute: max, Temperature: 1, Output limit: 128,000 tokens GPT-6 Astra: Reasoning: max, Output limit: 128,000 tokens	Higher measured value54.393 %	48.318 %	Vals AI Source details Data as of: 10/12/2026
Harvey's Legal Agent BenchmarkAntitrust Competition · Criteria passed Test details Test setup not documented as identicalNo direct winnerVersion: 1Metric: criteria pass rateScoring: Higher is betterClaude Fable 5.1: Compute: max, Temperature: 1, Output limit: 128,000 tokens GPT-6 Astra: Reasoning: max, Output limit: 128,000 tokens	Higher measured value89.855 %	88.768 %	Vals AI Source details Data as of: 10/12/2026
IOIIOI 2024 Test details Test setup not documented as identicalNo direct winnerVersion: 2Metric: accuracyScoring: Higher is betterClaude Fable 5.1: Compute: max, Temperature: 1, Output limit: 128,000 tokens GPT-6 Astra: Reasoning: max, Output limit: 128,000 tokens	No direct winner100 %	No direct winner100 %	Vals AI Source details Data as of: 10/14/2026
Legal Research BenchAdministrative / Regulatory · All-pass Test details Test setup not documented as identicalNo direct winnerVersion: 1Metric: all-pass rateScoring: Higher is betterClaude Fable 5.1: Compute: max, Temperature: 1, Output limit: 128,000 tokens GPT-6 Astra: Reasoning: max, Output limit: 128,000 tokens	Higher measured value68.182 %	53.03 %	Vals AI Source details Data as of: 10/12/2026
LMArena Code Arena WebDevImage to webdev Test details No documented difference in the test setupNo clear leadVersion: webdevMetric: WebDev Arena Bradley-Terry ratingScoring: Higher is betterClaude Fable 5.1: Snapshot: claude-fable-5.1-max, Reasoning: max, Harness: Arena Code Arena WebDev GPT-6 Astra: Snapshot: gpt-6-astra-max, Reasoning: max, Harness: Arena Code Arena WebDevClaude Fable 5.1: Arena rating from 2,470 published pairwise votes.GPT-6 Astra: Arena rating from 2,739 published pairwise votes.	1,719.751 95% confidence interval: ±13.642The Arena score measures human preference rather than a fixed capability test. The confidence interval, Arena category, and model configuration are part of the result.	No clear lead1,731.395 95% confidence interval: ±13.296The Arena score measures human preference rather than a fixed capability test. The confidence interval, Arena category, and model configuration are part of the result.	LMArena Source details Data as of: 10/09/2026
LMArena Text, Style ControlBusiness, management and finance Test details No documented difference in the test setupNo clear leadVersion: style-control-v1Metric: style-controlled Bradley-Terry ratingScoring: Higher is betterClaude Fable 5.1: Snapshot: claude-fable-5.1-max, Reasoning: max, Harness: Arena Text, style control GPT-6 Astra: Snapshot: gpt-6-astra-max, Reasoning: max, Harness: Arena Text, style controlClaude Fable 5.1: Arena rating from 2,589 published pairwise votes.GPT-6 Astra: Arena rating from 2,061 published pairwise votes.	No clear lead1,485.231 95% confidence interval: ±11.847The Arena score measures human preference rather than a fixed capability test. The confidence interval, Arena category, and model configuration are part of the result.	1,472.34 95% confidence interval: ±13.336The Arena score measures human preference rather than a fixed capability test. The confidence interval, Arena category, and model configuration are part of the result.	LMArena Source details Data as of: 10/13/2026
LMArena Vision, Style ControlChinese Test details No documented difference in the test setupNo clear leadVersion: style-control-v1Metric: style-controlled Bradley-Terry ratingScoring: Higher is betterClaude Fable 5.1: Snapshot: claude-fable-5.1-max, Reasoning: max, Harness: Arena Vision, style control GPT-6 Astra: Snapshot: gpt-6-astra-max, Reasoning: max, Harness: Arena Vision, style controlClaude Fable 5.1: Arena rating from 260 published pairwise votes.GPT-6 Astra: Arena rating from 230 published pairwise votes.	1,340.8 95% confidence interval: ±36.169The Arena score measures human preference rather than a fixed capability test. The confidence interval, Arena category, and model configuration are part of the result.	No clear lead1,357.126 95% confidence interval: ±39.572The Arena score measures human preference rather than a fixed capability test. The confidence interval, Arena category, and model configuration are part of the result.	LMArena Source details Data as of: 10/09/2026
ProgramBenchAlmost Resolved Test details Test setup not documented as identicalNo direct winnerVersion: 1Metric: almost-resolved rateScoring: Higher is betterClaude Fable 5.1: Compute: max, Temperature: 1, Output limit: 128,000 tokens GPT-6 Astra: Reasoning: max, Output limit: 128,000 tokens	Higher measured value53.5 %	50 %	Vals AI Source details Data as of: 10/12/2026
SAGEAutomata Test details Test setup not documented as identicalNo direct winnerVersion: 1Metric: accuracyScoring: Higher is betterClaude Fable 5.1: Compute: max, Temperature: 1, Output limit: 30,000 tokens GPT-6 Astra: Reasoning: max, Output limit: 30,000 tokens	34.524 %	Higher measured value37.302 %	Vals AI Source details Data as of: 10/08/2026
SRE BenchCapability Score Test details Test setup not documented as identicalNo direct winnerVersion: 1Metric: capability scoreScoring: Higher is betterClaude Fable 5.1: Compute: max, Temperature: 1, Output limit: 128,000 tokens GPT-6 Astra: Reasoning: max, Output limit: 128,000 tokens	36.641 %	Higher measured value63.868 %	Vals AI Source details Data as of: 10/14/2026
SWE-bench Multimodalbpmn-js Test details Test setup not documented as identicalNo direct winnerVersion: 1Metric: accuracyScoring: Higher is betterClaude Fable 5.1: Compute: max, Temperature: 1, Output limit: 128,000 tokens GPT-6 Astra: Reasoning: max, Output limit: 128,000 tokens	Higher measured value79.245 %	62.264 %	Vals AI Source details Data as of: 10/13/2026
Tax Agent BenchControversy & Precedence Analysis Test details Test setup not documented as identicalNo direct winnerVersion: 1Metric: accuracyScoring: Higher is betterClaude Fable 5.1: Compute: max, Temperature: 1, Output limit: 128,000 tokens GPT-6 Astra: Reasoning: max, Output limit: 128,000 tokens	Higher measured value26.667 %	13.333 %	Vals AI Source details Data as of: 10/12/2026
Terminal-Bench 2.1Easy Test details Test setup not documented as identicalNo direct winnerVersion: 2.1Metric: accuracyScoring: Higher is betterClaude Fable 5.1: Compute: high, Temperature: 1, Output limit: 128,000 tokens GPT-6 Astra: Reasoning: max, Output limit: 128,000 tokens	No direct winner100 %	No direct winner100 %	Vals AI Source details Data as of: 09/28/2026
Terminal-Bench 4.0Hardware Test details Test setup not documented as identicalNo direct winnerVersion: 4.0Metric: accuracyScoring: Higher is betterClaude Fable 5.1: Compute: max, Temperature: 1, Output limit: 128,000 tokens GPT-6 Astra: Reasoning: max, Output limit: 128,000 tokens	No direct winner60 %	No direct winner60 %	Vals AI Source details Data as of: 10/13/2026
Terminal-Bench ScienceEarth Sciences Test details Test setup not documented as identicalNo direct winnerVersion: 0.1Metric: accuracyScoring: Higher is betterClaude Fable 5.1: Compute: max, Harness: Mini-SWE-agent, Temperature: 1, Output limit: 128,000 tokens GPT-6 Astra: Reasoning: max, Harness: Mini-SWE-agent, Output limit: 128,000 tokens	25 %	Higher measured value62.5 %	Vals AI Source details Data as of: 10/14/2026
Time Horizon Index: KSPProgress efficiency Test details Test setup not documented as identicalNo direct winnerVersion: 1Metric: progressScoring: Higher is betterClaude Fable 5.1: Compute: max, Temperature: 1, Output limit: 128,000 tokens GPT-6 Astra: Reasoning: max, Output limit: 128,000 tokens	33.384 %	Higher measured value70.112 %	Vals AI Source details Data as of: 10/07/2026
Vals IndexCode Migration Test details Test setup not documented as identicalNo direct winnerVersion: 2.1Metric: benchmark scoreScoring: Higher is betterClaude Fable 5.1: Compute: max, Temperature: 1, Output limit: 128,000 tokens GPT-6 Astra: Reasoning: max, Output limit: 128,000 tokens	60.033 %	Higher measured value68.462 %	Vals AI Source details Data as of: 10/13/2026
Code MigrationCOBOL Test details Test setup not documented as identicalNo direct winnerVersion: 1Metric: accuracyScoring: Higher is betterClaude Fable 5.1: Compute: max, Temperature: 1, Output limit: 128,000 tokens GPT-6 Astra: Reasoning: max, Output limit: 128,000 tokens	60 %	Higher measured value68.571 %	Vals AI Source details Data as of: 10/12/2026
CyberBench v1.1PoC Test details Test setup not documented as identicalNo direct winnerVersion: 1.1Metric: accuracyScoring: Higher is betterClaude Fable 5.1: Compute: max, Temperature: 1, Output limit: 128,000 tokens GPT-6 Astra: Reasoning: max, Output limit: 128,000 tokens	Higher measured value53.333 %	0 %	Vals AI Source details Data as of: 10/12/2026
EMBDataroom Test details Test setup not documented as identicalNo direct winnerVersion: 1Metric: accuracyScoring: Higher is betterClaude Fable 5.1: Compute: max, Temperature: 1, Output limit: 128,000 tokens GPT-6 Astra: Reasoning: max, Output limit: 128,000 tokens	Higher measured value86.736 %	79.319 %	Vals AI Source details Data as of: 10/12/2026
Finance Agent (v2)All-Pass Test details Test setup not documented as identicalNo direct winnerVersion: 2Metric: accuracyScoring: Higher is betterClaude Fable 5.1: Compute: max, Temperature: 1, Output limit: 128,000 tokens GPT-6 Astra: Reasoning: max, Output limit: 128,000 tokens	Higher measured value47.773 %	40.252 %	Vals AI Source details Data as of: 10/12/2026
Key to the marks:better valuebetter, but not conclusivetie

Our model tests

Our tests

SVG short test

One SVG prompt and fixed output settings. Captures are shown only with 0 reported reasoning tokens or documented lack of reasoning controls. Identical test conditions do not make the models' internal compute identical.

Shared test round: 4ae5d8eeb9a9

Claude Fable 5.1

Unavailable in this round

GPT-6 Astra

Unavailable in this round

Earlier tests

Pelican on a bicycle

A community classic for free-form SVG drawing.

Claude Fable 5.1

Reasoning (requested): high

Current attempt

Run date: Sep 7, 2026
API provider: openrouter
Requested API model: anthropic/claude-fable-5.1
API model reported in response: anthropic/claude-fable-5.1
Reasoning setting: high
Pinned endpoint: Anthropic | anthropic/claude-fable-5.1-20260831
Temperature: Provider default
Token limit including reasoning: 8,192 tokens
Test protocol: community-visual-tests-v1
Technical checks: Passed, not a quality rating
Response time: 27 s

Visually reviewed. This individual test does not establish overall model quality.

GPT-6 Astra

Reasoning (requested): high

Current attempt

Run date: Sep 7, 2026
API provider: openrouter
Requested API model: openai/gpt-6-astra
API model reported in response: openai/gpt-6-astra
Reasoning setting: high
Pinned endpoint: OpenAI | openai/gpt-6-astra-20260903
Temperature: Provider default
Token limit including reasoning: 8,192 tokens
Test protocol: community-visual-tests-v1
Technical checks: Passed, not a quality rating
Response time: 141.7 s

Visually reviewed. This individual test does not establish overall model quality.

Prompt and test conditions

Generate an SVG of a pelican riding a bicycle

Token limit including reasoning: 8,192 tokens

Task origin (Simon Willison)

How to read this comparison

Winning one benchmark is not an overall verdict

Your workload, budget, and required context length matter most. A coding benchmark says little about visual understanding or agent performance.

The standard price table uses direct developer rates. The provider comparison labels OpenRouter endpoints, regions, and special tariffs separately.

Every result links to its measurement source and retrieval date. Labels distinguish vendor reports, official benchmark leaderboards, and independent evaluations. Disputed or archived results are excluded.

More matchups

Compare Claude Fable 5.1 and GPT-6 Astra with other leading models using the same data and benchmark logic.

View the complete LLM comparison

Claude Fable 5.1 vs. GPT-6 Astra

Change model selection

General specifications

API pricing compared

Shared benchmarks

Capability profile from matched benchmarks

Our tests

SVG short test

Claude Fable 5.1

GPT-6 Astra

Pelican on a bicycle

Winning one benchmark is not an overall verdict

Related LLM comparisons