LLM Reference

Claude Fable 5 vs Claude Opus 4.8

Claude Fable 5 and Claude Opus 4.8 are Anthropic's most important in-family comparison for teams deciding whether to move from the prior Opus generation to the Mythos-class Fable route. Anthropic says Fable 5's capabilities exceed any model it had previously made generally available, but Fable 5 also includes safeguards that can route some requests differently than Opus 4.8.

Pick Claude Fable 5 when you need Anthropic's most capable widely released Mythos-class route and can accept documented refusal and fallback behavior after verifying live access. Keep Claude Opus 4.8 when you need the prior Opus behavior profile, fallback predictability, or a route that may still be easier to reason about for some regulated workflows. Treat the June 9 launch, June 12 suspension, and July 1 redeploy timeline as part of the product decision, not a footnote.

Decision scorecard

Local evidence first
SignalClaude Fable 5Claude Opus 4.8
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$25/1M tokens
Provider routes8 tracked6 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 7.4 points.
  • Claude Fable 5 has broader tracked provider coverage for fallback and route flexibility.
  • Local decision data tags Claude Fable 5 for Coding, RAG, and Agents.
Choose Claude Opus 4.8 when...
  • Claude Opus 4.8 has the lower cheapest tracked output price at $25/1M tokens.
  • Claude Opus 4.8 uniquely exposes Code execution, Computer use, and Parallel agents in local model data.
  • Local decision data tags Claude Opus 4.8 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 Opus 4.8

Claude Fable 5

$20,500

Cheapest tracked route/tier: Anthropic

Claude Opus 4.8

$10,250

Cheapest tracked route/tier: Anthropic

Estimated monthly gap: $10,250. Batch, cache, alternate speed tiers, and negotiated pricing are excluded from this local estimate.

Switch friction

Claude Fable 5 -> Claude Opus 4.8
  • Provider overlap exists on Anthropic, AWS Bedrock, and GCP Vertex AI; start route-level A/B tests there.
  • Claude Opus 4.8 is $25/1M tokens lower on cheapest tracked output pricing before cache, batch, or negotiated discounts.
  • Claude Opus 4.8 adds Code execution, Computer use, and Parallel agents in local capability data.
Claude Opus 4.8 -> 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 $25/1M tokens higher on cheapest tracked output pricing, so quality gains need to justify the spend.
  • Check replacement coverage for Code execution, Computer use, and Parallel agents before moving production traffic.

Specs

Specification
Released2026-06-092026-05-28
Context window1m1m
Parameters
ArchitectureDecoder OnlyDecoder Only
LicenseProprietaryProprietary
OpennessProprietaryProprietary
WeightsNot releasedNot released
CodeUnknownNot released
Commercial useCommercial use: conditionalCommercial use: conditional
Knowledge cutoff-2026-01

Pricing and availability

Pricing attributeClaude Fable 5Claude Opus 4.8
Input price$10/1M tokens$5/1M tokens
Output price$50/1M tokens$25/1M tokens
Providers

Capabilities

CapabilityClaude Fable 5Claude Opus 4.8
VisionYesYes
MultimodalYesYes
ReasoningYesYes
JSON / Tool useYesYes
Structured outputsYesYes
Code executionNoYes
IDE integrationNoNo
Computer useNoYes
Parallel agentsNoYes

Benchmarks

BenchmarkClaude Fable 5Claude Opus 4.8
SWE-bench Verified96.088.6
SWE-bench Pro80.369.2
OSWorld-Verified85.083.4
GDPval-AA1932.01890.0
CursorBench72.953.1
CursorBench70.553.1
CursorBench68.453.1
CursorBench66.553.1
CursorBench65.253.1
CursorBench62.153.1

Continue comparing

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