LLM Reference

Claude Fable 5 vs Claude Sonnet 5

Claude Fable 5 and Claude Sonnet 5 are both Anthropic Claude 5 models, but they target different buyer decisions. Fable 5 is Anthropic's most capable widely released Mythos-class model for demanding reasoning and long-horizon agentic work, with documented refusal/fallback behavior and $10/M input pricing. Sonnet 5 is the balanced Sonnet-tier route for day-to-day coding and workflow automation at lower token cost.

Pick Claude Fable 5 when the workload needs Anthropic's highest generally available capability tier and you can accept refusal/fallback behavior plus the $10/M input and $50/M output launch price once access is restored. Pick Claude Sonnet 5 for cost-sensitive production coding, higher throughput, and mainstream API adoption where Sonnet-tier capability is enough.

Decision scorecard

Local evidence first
SignalClaude Fable 5Claude Sonnet 5
Best forreasoning-heavy apps, multimodal apps, and tool-calling agentsreasoning-heavy apps, multimodal apps, and tool-calling agents
Decision fitCoding, RAG, and AgentsCoding, RAG, and Agents
Context window1m1m
Cheapest output$50/1M tokens$10/1M tokens
Provider routes8 tracked5 tracked
Shared benchmarksSWE-bench Verified leader10 shared

Decision tradeoffs

Choose Claude Fable 5 when...
  • Claude Fable 5 holds a shared-benchmark lead on SWE-bench Verified, ahead by 10.8 points.
  • Claude Fable 5 has broader tracked provider coverage for fallback and procurement flexibility.
  • Local decision data tags Claude Fable 5 for Coding, RAG, and Agents.
Choose Claude Sonnet 5 when...
  • Claude Sonnet 5 holds a shared-benchmark lead on GDP.pdf, ahead by 51.8 points.
  • Claude Sonnet 5 has the lower cheapest tracked output price at $10/1M tokens.
  • Claude Sonnet 5 uniquely exposes Code execution in local model data.
  • Local decision data tags Claude Sonnet 5 for Coding, RAG, and Agents.

Monthly cost at traffic

Estimate token spend from the cheapest tracked input and output route or tier on this page.

Lower estimate Claude Sonnet 5

Claude Fable 5

$20,500

Cheapest tracked route/tier: Anthropic

Claude Sonnet 5

$4,100

Cheapest tracked route/tier: Anthropic

Estimated monthly gap: $16,400. Batch, cache, alternate speed tiers, and negotiated pricing are excluded from this local estimate.

Switch friction

Claude Fable 5 -> Claude Sonnet 5
  • Provider overlap exists on Anthropic, AWS Bedrock, and GCP Vertex AI; start route-level A/B tests there.
  • Claude Sonnet 5 is $40/1M tokens lower on cheapest tracked output pricing before cache, batch, or negotiated discounts.
  • Claude Sonnet 5 adds Code execution in local capability data.
Claude Sonnet 5 -> Claude Fable 5
  • Provider overlap exists on Anthropic, AWS Bedrock, and GCP Vertex AI; start route-level A/B tests there.
  • Claude Fable 5 is $40/1M tokens higher on cheapest tracked output pricing, so quality gains need to justify the spend.
  • Check replacement coverage for Code execution before moving production traffic.

Specs

Specification
Released2026-06-092026-06-30
Context window1m1m
Parameters
ArchitectureDecoder OnlyDecoder Only
LicenseProprietaryProprietary
OpennessProprietaryProprietary
WeightsNot releasedNot released
CodeUnknownUnknown
Commercial useCommercial use: conditionalCommercial use: conditional
Knowledge cutoff-2026-01

Pricing and availability

Pricing attributeClaude Fable 5Claude Sonnet 5
Input price$10/1M tokens$2/1M tokens
Output price$50/1M tokens$10/1M tokens
Providers

Capabilities

CapabilityClaude Fable 5Claude Sonnet 5
VisionYesYes
MultimodalYesYes
ReasoningYesYes
Function callingYesYes
Tool useYesYes
Structured outputsYesYes
Code executionNoYes
IDE integrationNoNo
Computer useNoNo
Parallel agentsNoNo

Benchmarks

BenchmarkClaude Fable 5Claude Sonnet 5
SWE-bench Verified96.085.2
SWE-bench Pro80.363.2
OSWorld-Verified85.081.2
GDP.pdf29.881.6
CursorBench72.947.7
CursorBench70.547.7
CursorBench68.447.7
CursorBench66.547.7
CursorBench65.247.7
CursorBench62.147.7

Deep dive

The tier split is the main decision frame. Fable 5 is positioned above Sonnet 5 for long-horizon agentic engineering, computer use, and hard reasoning, while Sonnet 5 is the practical balanced model for teams that do not need Fable 5's premium tier or safety-routing complexity.

Fable 5's refusal and fallback behavior is unique in this pair. Anthropic documents safety classifiers that may refuse or route certain requests to Claude Opus 4.8. Sonnet 5 should not inherit those specs automatically; treat Fable-specific integration behavior as a Fable-only concern.

Availability still matters. Fable 5 launched June 9, 2026, was suspended June 12, and Anthropic says it redeployed July 1, 2026. Verify live Fable 5 access on your provider before production; Sonnet 5 remains the safer default Claude route when you need confirmed availability today.

Do not collapse Fable 5 and Claude Mythos 5. Mythos 5 is a restricted-access Glasswing model with different availability and starred benchmark rows. This page compares two generally marketed Claude tiers, not Mythos access programs.

FAQ

When should I use Claude Fable 5 instead of Sonnet 5?

Use Fable 5 when you need Anthropic's highest generally available capability tier for long-horizon agentic work and can accept the higher $10/M input price plus refusal/fallback behavior once access is restored. Use Sonnet 5 for balanced coding and workflow automation at lower cost.

How does Fable 5 refusal handling affect integrations?

Fable 5 may refuse or fall back to Claude Opus 4.8 on certain requests because of safety classifiers. Sonnet 5 does not share that exact behavior profile, so do not assume Sonnet integrations will see the same routing pattern.

Which Claude model is best for long-horizon work?

Claude Fable 5 is Anthropic's most capable widely released model for demanding long-horizon agentic work in the current taxonomy, but verify live access before committing. Sonnet 5 is the better default when cost and availability matter more than maximum capability tier.

Is Claude Fable 5 available now?

Anthropic launched Fable 5 June 9, suspended access June 12, and says it redeployed July 1, 2026. Verify your Claude API or cloud partner route before production. Sonnet 5 remains the safer default when you need confirmed availability today.

Continue comparing

Last reviewed: 2026-07-26. Data sourced from public model cards and provider documentation.