Model comparison
Claude Fable 5 vs GPT-5.6 Sol
Claude Fable 5 is the stronger model overall, scoring 66.8 to 65.0 on the Noometry Index. GPT-5.6 Sol costs 2.5× less per token, which makes it the better buy when Claude Fable 5's lead doesn't matter for your workload.
Last verified . 61 shared benchmarks.
Summary
- They share 61 benchmarks with published results for both. Claude Fable 5 scores higher in 8 categories and GPT-5.6 Sol in 2 categories; 8 gaps are clear of the uncertainty.
- The widest gap is in coding, where Claude Fable 5 leads 70.6 to 65.1.
- The biggest single-benchmark swing is MirrorCode: 63.9% for Claude Fable 5 and 20% for GPT-5.6 Sol.
- GPT-5.6 Sol is cheaper at $4 / $20 per million input/output tokens, against $10 / $50 for Claude Fable 5.
- GPT-5.6 Sol accepts more context: 1.05M tokens versus 1M.
Side by side
| Claude Fable 5 | GPT-5.6 Sol | |
|---|---|---|
| Provider | Anthropic | OpenAI |
| Noometry Index | 66.8 | 65.0 |
| Released | 2026-06-07 | 2026-07-09 |
| Weights | Proprietary | Proprietary |
| Context window | 1M | 1.05M |
| Max output | 128K | 128K |
| Input $ / M tokens | $10 | $4 |
| Output $ / M tokens | $50 | $20 |
| Results tracked | 62 | 65 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding Claude Fable 5 leads
Claude Fable 5: 70.6 (#4), GPT-5.6 Sol: 65.1 (#7)
| Benchmark | Claude Fable 5 | GPT-5.6 Sol |
|---|---|---|
| DeepSWE | 69.9% | 72.7% |
| FrontierCode | 53.5% | 47.5% |
| LMArena WebDev | 1625 | 1618 |
| FrontierSWE | 47% | 32.2% |
| SciCode | 61% | 57.1% |
| GSO | 78.4% | 76.5% |
| WeirdML | 91.9% | 89.4% |
| LMArena Coding | 1519 | 1498 |
| MirrorCode | 63.9% | 20% |
| ALE-Bench | 2,041 | 2,177 |
| CursorBench | — | 41.7% |
Agentic & Tool Use Claude Fable 5 leads
Claude Fable 5: 54.0 (#2), GPT-5.6 Sol: 50.3 (#7)
| Benchmark | Claude Fable 5 | GPT-5.6 Sol |
|---|---|---|
| APEX-Agents | 63.6% | 51.4% |
| τ²-bench Banking | 39.7% | 46.9% |
| PostTrainBench | 41.8% | 36.2% |
| GBAEval | 74.5% | 52.6% |
| GDP.pdf | 30% | 30.7% |
| LMArena Search | 1230 | 1257 |
| Vending-Bench 2 | 5,680 | 9,619 |
| OSWorld 2.0 | — | 27.3% |
| Remote Labor Index | 16.1% | — |
| BALROG | — | 60% |
Reasoning Claude Fable 5 leads
Claude Fable 5: 76.8 (#6), GPT-5.6 Sol: 74.8 (#8)
| Benchmark | Claude Fable 5 | GPT-5.6 Sol |
|---|---|---|
| ARC-AGI-2 | 89.2% | 92.5% |
| SimpleBench | 81.9% | 71.7% |
| Kagi LLM Benchmark | 91.4% | 67% |
| NYT Connections (extended) | 92.7% | 93.8% |
| ARC-AGI-1 | 98.5% | 97.5% |
| CritPt | 28.6% | 32.3% |
| Chess Puzzles | 41% | 64% |
| EnigmaEval | 39.3% | 37.1% |
| EBR-Bench | 39.5% | 44.8% |
| LMArena Hard Prompts | 1508 | 1484 |
| Mystery Game Puzzles | 52% | 58% |
| DTBench | 98.4% | 96% |
| LMCA | 61.1% | 59.2% |
| Surface Evolver Bench | 95% | 93.1% |
| Bench to the Future 3 | 0.13 | 0.14 |
| Epoch Capabilities Index | 162.06 | 161.66 |
Math Claude Fable 5 leads
Claude Fable 5: 88.5 (#5), GPT-5.6 Sol: 85.6 (#9)
| Benchmark | Claude Fable 5 | GPT-5.6 Sol |
|---|---|---|
| FrontierMath (Tiers 1-3) | 87% | 89.1% |
| FrontierMath Tier 4 | 90.2% | 82.9% |
| OTIS Mock AIME 2024-2025 | 100% | 100% |
| ProofBench | 95% | 83% |
| LMArena Math | 1519 | 1474 |
| FrontierMath Erdős | 0% | 0% |
Knowledge GPT-5.6 Sol leads
Claude Fable 5: 62.2 (#25), GPT-5.6 Sol: 64.3 (#18)
| Benchmark | Claude Fable 5 | GPT-5.6 Sol |
|---|---|---|
| GPQA Diamond | 85.9% | 93.5% |
| SimpleQA Verified | 70.7% | 69.7% |
| LMArena Expert | 1534 | 1516 |
| Vectara Hallucination Rate | — | 12.4% |
Multimodal GPT-5.6 Sol leads
Claude Fable 5: 45.3 (#17), GPT-5.6 Sol: 48.6 (#9)
| Benchmark | Claude Fable 5 | GPT-5.6 Sol |
|---|---|---|
| LMArena Vision | 1324 | 1281 |
| Blueprint-Bench 2 | 38.6% | 33.6% |
| Furniture Assembly | 35.8% | 56.7% |
| LMArena Document | 1496 | 1483 |
Multilingual Claude Fable 5 leads
Claude Fable 5: 57.3 (#9), GPT-5.6 Sol: 55.3 (#32)
| Benchmark | Claude Fable 5 | GPT-5.6 Sol |
|---|---|---|
| LMArena Non-English | 1481 | 1452 |
| LMArena Chinese | 1543 | 1527 |
| LMArena French | 1505 | 1477 |
| LMArena German | 1486 | 1476 |
| LMArena Japanese | 1506 | 1471 |
| LMArena Korean | 1488 | 1442 |
| LMArena Russian | 1504 | 1468 |
| LMArena Spanish | 1498 | 1441 |
Instruction Following Too close to call
Claude Fable 5: 78.6 (#8), GPT-5.6 Sol: 77.7 (#16)
| Benchmark | Claude Fable 5 | GPT-5.6 Sol |
|---|---|---|
| LMArena Instruction Following | 1502 | 1482 |
Long Context Too close to call
Claude Fable 5: 46.3 (#23), GPT-5.6 Sol: 45.4 (#42)
| Benchmark | Claude Fable 5 | GPT-5.6 Sol |
|---|---|---|
| LMArena Longer Query | 1509 | 1480 |
Writing & Preference Claude Fable 5 leads
Claude Fable 5: 75.9 (#5), GPT-5.6 Sol: 73.3 (#12)
| Benchmark | Claude Fable 5 | GPT-5.6 Sol |
|---|---|---|
| LMArena Text | 1491 | 1457 |
| LMArena Creative Writing | 1494 | 1448 |
| EQ-Bench Creative Writing | 1943 | 1972 |
| EQ-Bench 4 | 1340 | 1250 |
| LMArena Multi-Turn | 1504 | 1460 |
Frequently asked questions
Is Claude Fable 5 better than GPT-5.6 Sol?
Claude Fable 5 is the stronger model overall, scoring 66.8 to 65.0 on the Noometry Index. GPT-5.6 Sol costs 2.5× less per token, which makes it the better buy when Claude Fable 5's lead doesn't matter for your workload.
Which is cheaper, Claude Fable 5 or GPT-5.6 Sol?
GPT-5.6 Sol is cheaper. It lists at $4 per million input tokens and $20 per million output tokens; Claude Fable 5 lists at $10 and $50.
Is Claude Fable 5 or GPT-5.6 Sol better for coding?
Claude Fable 5 scores higher on coding benchmarks: 70.6 versus 65.1 in the Noometry coding category.
Which has the bigger context window?
GPT-5.6 Sol does, with 1.05M tokens against 1M.
How many benchmarks do Claude Fable 5 and GPT-5.6 Sol share?
61 benchmarks have published results for both models. Claude Fable 5 has 62 scored results on Noometry and GPT-5.6 Sol has 65.