Claude Fable 5 vs Claude Opus 4.7
Different efforts: max · high · Compare all at high
Humanity's Last Exam Humanity's Last Exam: Very hard expert questions across many fields, written to stump today’s models
Fable 555.5
Opus 4.733.3
Terminal-Bench 4.0 Terminal-Bench 4.0 · source setting 2: Agents use a terminal to complete executable coding, system administration and research tasks in the 4.0 suite
Fable 542.4
Opus 4.7—
AutomationBench-AA AutomationBench-AA: Agents orchestrate SaaS workflows through APIs; objectives receive credit only when the task has no guardrail violation
Fable 554.1
Opus 4.7—
Arena Text · creative writing Arena Text · creative writing: Human pairwise preferences in the stated text vote pool
Fable 5 · high1,501
Opus 4.71,489
MMMU-Pro (vision) MMMU-Pro (vision): College-level questions that require understanding images: charts, diagrams, photos
Fable 5—
Opus 4.776.4
| Benchmark | Fable 5 | Opus 4.7 |
|---|---|---|
| Humanity's Last Exam | 55.5 | 33.3 |
| Terminal-Bench 4.0 | 42.4 | — |
| AutomationBench-AA | 54.1 | — |
| Arena Text · creative writing | 1,501 · high | 1,489 |
| MMMU-Pro (vision) | — | 76.4 |
Specs
| Spec | Fable 5 | Opus 4.7 |
|---|---|---|
| Input, per 1M | $10 | $5 |
| Output, per 1M | $50 | $25 |
| Cost per task? | $8.75 | — |
| Speed, tok/s | 67 | 59 |
| Context | 1M | 1M |
| Weights | Closed | Closed |
| Released | Jun 2026 | Apr 2026 |
| See images | ✓ | ✓ |
| Read PDFs | Partial | Partial |
| Hear audio | — | — |
| Watch video | — | — |
| Call tools | ✓ | ✓ |
| Force a tool call | Partial | Partial |
| Return strict JSON | ✓ | ✓ |
| Search the web | ? | ? |
| Run code | ? | ? |
| Use a computer | ? | ? |
What breaks when you switch
Fable 5 → Opus 4.7
Nothing known.
Opus 4.7 → Fable 5
Nothing known.