Claude Fable 5 vs Claude Opus 4.8 vs GPT-5.5 vs Gemini 3.1 Pro

Best cell per row highlighted. Null means undisclosed — never counted as zero.

Anthropic
Claude Fable 5GA
9.5
AI Panel
Anthropic
Claude Opus 4.8GA
9.2
AI Panel
OpenAI
GPT-5.5GA
8.7
AI Panel
Google
Gemini 3.1 Propreview
8.7
AI Panel

Identity & lifecycle

ProviderAnthropicAnthropicOpenAIGoogle
Family / tierClaude 5 · FableClaude 4 · OpusGPT-5 · ProGemini 3 · Pro
StatusGAGAGApreview
Released2026-06-092026-05-282026-04-242026-02-19
Knowledge cutoff2026-012026-012025-122025-01

Architecture & context

Context window1M1M1.1M1.0M
Max output tokens128K128K128K66K
Input modalitiestext, image, filetext, imagetext, imagetext, image, audio, video
Reasoning modealwaysalwaysoptionaloptional
Open weightsNoNoNoNo

Pricing (per Mtok)

Input$10$5$5$2
Output$50$25$30$12
Cached input$1$0.5$0.5$0.2
Batch input$5$2.5$2.5$1
Free tierNoNoNoNo

Speed

Output speed (tok/s)—6068.7142.7
Time to first token (s)—0.56430.78
Latency tierslowmediumslowslow

Trust & deployment

Trains on inputsNoNoNoNo
LicenseProprietaryProprietaryProprietaryProprietary
CloudsBedrock, Vertex AI, Azure AI FoundryBedrock, Vertex AI, Azure AI FoundryAzure OpenAI, Azure AI FoundryVertex AI, GCP
ComplianceSOC2, HIPAA, GDPR, ISO27001SOC2, HIPAA, GDPR, ISO27001SOC2, GDPR, CCPA, HIPAASOC2, HIPAA, GDPR, ISO27001, FedRAMP, CCPA

AI Panel scoring

Unified score9.59.28.78.7
Decision Maker9.5999
Domain Strategist9.598.89
Finance Lead78.578.5
Domain Practitioner9.59.59.28.5
Power User9998.5
Skeptic7.587.87.5
Value score6.586.58.5

Benchmarks

MMLU———92.6
MMLU-Pro——92.4—
GPQA Diamond—93.693.694.3
SWE-bench Verified9588.688.780.6
LiveCodeBench———2887
MMMU——81.280.5
TAU-bench——9890.8
Terminal-Bench8874.682.768.5
MRCR Long Context——7484.9
Humanity's Last Exam——41.444.4
LMArena Elo1506148114821487
LMArena Coding Elo1552153315201520

Scores link to their sources. Missing cells mean the vendor hasn't disclosed a result — honesty over padding.