Claude Fable 5 vs GPT-5.6 Sol
Claude Fable 5 leads GPT-5.6 Sol by 2.3 points on the Shogo Agent Index, but their ranges overlap, so treat it as close. Claude Fable 5 wins 18 and GPT-5.6 Sol wins 6 of 25 shared benchmarks. GPT-5.6 Sol is cheaper for agent workloads.
Add more models to this comparison →Pillars
| Pillar | Claude Fable 5 | GPT-5.6 Sol |
|---|---|---|
| Agentic coding | 74 | 71 |
| Long-horizon autonomy | 69 | 72 |
| Knowledge-work agents | 52 | 46 |
| Computer use | 55 | 53 |
| Security & context learning | 59 | 58 |
Shared benchmarks
| Benchmark | Claude Fable 5 | GPT-5.6 Sol |
|---|---|---|
| APEX-Agentsindex | 63.6% | 51.4% |
| DeepSWEindex | 69.9% | 72.7% |
| Vending-Bench 2index | $5,680 | $9,619 |
| FrontierSWEindex | 47.0% | 32.2% |
| GSO-Benchindex | 78.4% | 76.5% |
| PostTrainBenchindex | 37.3% | 31.0% |
| ARC-AGI | 98.5% | 97.5% |
| ARC-AGI-2 | 89.2% | 92.5% |
| Chess Puzzles | 47.0% | 68.0% |
| CritPt | 28.6% | 32.3% |
| DTBench | 98.4% | 96.0% |
| FrontierCode | 53.5% | 47.5% |
| FrontierMath Tier 4 | 90.2% | 82.9% |
| FrontierMath Tiers 1-3 | 87.0% | 89.1% |
| GeoBench | 89.0% | 73.0% |
| GPQA Diamond | 93.9% | 93.5% |
| LMCA | 61.2% | 59.2% |
| MirrorCode | 70.0% | 20.0% |
| OTIS Mock AIME | 100.0% | 100.0% |
| ProofBench | 95.0% | 83.0% |
| SciCode | 61.0% | 57.8% |
| SimpleBench | 81.9% | 71.7% |
| SimpleQA Verified | 70.7% | 69.7% |
| WebDev Arena | 1628 | 1617 |
| WeirdML v3 | 15.6% | 15.0% |
Specs and price
| Claude Fable 5 | GPT-5.6 Sol | |
|---|---|---|
| Agent Index | 64.1 (#12) | 61.8 (#14) |
| Agent $/M | $4.62 | $1.85 |
| Input / output $/M | $10.00 / $50.00 | $4.00 / $20.00 |
| Context | 1M | 1.1M |
| Released | Jun 9, 2026 | Aug 7, 2026 |
| Weights | Closed | Closed |
| API hosts | 4 | 5 |
Don’t pick one. Route.
Shogo can run Claude Fable 5 and GPT-5.6 Sol side by side and send each step of an agent to the one that’s cheaper for it.
Try both in Shogo