Claude Fable 5.1 vs Claude Haiku 5.5
Claude Fable 5.1 leads Claude Haiku 5.5 by 10.7 points on the Shogo Agent Index, but their ranges overlap, so treat it as close. Claude Fable 5.1 wins 20 and Claude Haiku 5.5 wins 1 of 21 shared benchmarks. Claude Haiku 5.5 is cheaper for agent workloads.
Add more models to this comparison →Pillars
| Pillar | Claude Fable 5.1 | Claude Haiku 5.5 |
|---|---|---|
| Agentic coding | 80 | 67 |
| Long-horizon autonomy | 71 | 62 |
| Knowledge-work agents | 55 | 46 |
| Computer use | 62 | 47 |
| Security & context learning | 61 | 55 |
Shared benchmarks
| Benchmark | Claude Fable 5.1 | Claude Haiku 5.5 |
|---|---|---|
| APEX-Agentsindex | 68.6% | 56.1% |
| Vending-Bench 2index | $5,422 | $3,602 |
| FrontierSWEindex | 56.3% | 43.8% |
| CursorBenchindex | 51.8% | 48.4% |
| ARC-AGI | 97.5% | 89.0% |
| ARC-AGI-2 | 90.0% | 61.3% |
| Chess Puzzles | 48.0% | 55.0% |
| CritPt | 31.1% | 22.6% |
| DTBench | 97.6% | 95.0% |
| FrontierCode | 50.9% | 46.4% |
| FrontierMath Tier 4 | 87.8% | 46.3% |
| FrontierMath Tiers 1-3 | 90.2% | 75.1% |
| Humanity's Last Exam | 46.5% | 30.1% |
| LMCA | 65.5% | 51.1% |
| MirrorCode | 93.3% | 78.9% |
| OTIS Mock AIME | 100.0% | 98.9% |
| ProofBench | 100.0% | 89.0% |
| SciCode | 63.1% | 55.0% |
| SimpleBench | 86.6% | 59.2% |
| SimpleQA Verified | 70.8% | 23.8% |
| WeirdML v3 | 26.0% | 11.7% |
Specs and price
| Claude Fable 5.1 | Claude Haiku 5.5 | |
|---|---|---|
| Agent Index | 68.2 (#10) | 57.5 (#18) |
| Agent $/M | $4.01 | $0.050 |
| Input / output $/M | $10.00 / $50.00 | $0.10 / $0.50 |
| Context | 1M | 1M |
| Released | Sep 1, 2026 | Oct 7, 2026 |
| Weights | Closed | Closed |
| API hosts | 3 | 5 |
Don’t pick one. Route.
Shogo can run Claude Fable 5.1 and Claude Haiku 5.5 side by side and send each step of an agent to the one that’s cheaper for it.
Try both in Shogo