Claude Fable 5 vs Claude Opus 4.6
EBR Bench?
Fable 539.5
Opus 4.6at max12.7
MCP-Atlas Public?
Fable 5—
Opus 4.6at max79
MCP-Atlas · full 1000?
Fable 5—
Opus 4.6at max76.8
| Benchmark | Fable 5 | Opus 4.6 |
|---|---|---|
| AutomationBench-AA | 54.1 | — |
| τ³-bench · banking · 1.0.1 | 39.7 | 27.3 · max |
| Terminal-Bench Hard (version unstated) | 62.9 | 48.5 |
| AA-LCR (long context) 1.1 | 82.3 | 67 |
| DeepSWE 1.1 | 69.7 | 31.9 · max |
| EBR Bench | 39.5 | 12.7 · max |
| Furniture Assembly | 35.8 | 28.3 · max |
| GDPval-AA (work tasks) 2.1 | 55.5 | — |
| MCP-Atlas Public | — | 79 · max |
| MCP-Atlas · full 1000 | — | 76.8 · max |
| τ³-Banking · AA | 38.1 | — |
| AA Analyst Agent | 48.8 | — |
| APEX-Agents | — | 33 · max |
| Terminal-Bench 4.0 | 42.4 | — |
| AA-Briefcase 1.1 · Elo | 1,539.3 | — |
| EnterpriseOps-Gym-AA | 51.1 | — |
| SWE-bench Pro (private) | — | 47.1 |
Specs
| Spec | Fable 5 | Opus 4.6 |
|---|---|---|
| Input, per 1M | $10 | $5 |
| Output, per 1M | $50 | $25 |
| Cost per task? | $8.75 | — |
| Speed, tok/s | 67 | 45 |
| Context | 1M | 1M |
| Weights | Closed | Closed |
| Released | Jun 2026 | Feb 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 4.6
Nothing known.
Opus 4.6 → Fable 5
Nothing known.