Model comparison
Claude Fable 5 vs Claude Opus 4.8
Claude Fable 5 is the stronger model overall, scoring 66.3 to 60.7 on the Noometry Index. Claude Opus 4.8 costs 2.0× less per token, which makes it the better buy when Claude Fable 5's lead doesn't matter for your workload.
Last verified . 58 shared benchmarks.
Summary
- They share 58 benchmarks with published results for both. Claude Fable 5 scores higher in 10 categories and Claude Opus 4.8 in 0 categories; 8 gaps are clear of the uncertainty.
- The widest gap is in reasoning, where Claude Fable 5 leads 76.8 to 64.6.
- The biggest single-benchmark swing is FrontierMath Tier 4: 90.2% for Claude Fable 5 and 56.1% for Claude Opus 4.8.
- Claude Opus 4.8 is cheaper at $5 / $25 per million input/output tokens, against $10 / $50 for Claude Fable 5.
Side by side
| Claude Fable 5 | Claude Opus 4.8 | |
|---|---|---|
| Provider | Anthropic | Anthropic |
| Noometry Index | 66.3 | 60.7 |
| Released | 2026-06-07 | 2026-05-28 |
| Weights | Proprietary | Proprietary |
| Context window | 1M | 1M |
| Max output | 128K | 128K |
| Input $ / M tokens | $10 | $5 |
| Output $ / M tokens | $50 | $25 |
| Results tracked | 61 | 65 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding Claude Fable 5 leads
Claude Fable 5: 67.0 (#4), Claude Opus 4.8: 60.2 (#9)
| Benchmark | Claude Fable 5 | Claude Opus 4.8 |
|---|---|---|
| DeepSWE | 69.9% | 59% |
| FrontierCode | 53.5% | 46.5% |
| LMArena WebDev | 1626 | 1556 |
| SciCode | 61% | 53.5% |
| WeirdML | 91.9% | 82.9% |
| LMArena Coding | 1519 | 1491 |
| ALE-Bench | 2,041 | 1,564 |
| FrontierSWE | 47% | — |
| GSO | — | 47.1% |
| MirrorCode | 63.9% | — |
Agentic & Tool Use Claude Fable 5 leads
Claude Fable 5: 54.1 (#2), Claude Opus 4.8: 47.6 (#10)
| Benchmark | Claude Fable 5 | Claude Opus 4.8 |
|---|---|---|
| APEX-Agents | 63.6% | 48.9% |
| Remote Labor Index | 16.1% | 8.3% |
| τ²-bench Banking | 39.7% | 39.7% |
| PostTrainBench | 41.8% | 33.8% |
| GBAEval | 74.5% | 70.9% |
| GDP.pdf | 30% | 24% |
| LMArena Search | 1230 | 1204 |
| Vending-Bench 2 | 5,680 | 5,787 |
| OSWorld 2.0 | — | 20.6% |
| DeepResearch Bench | — | 50.2% |
Reasoning Claude Fable 5 leads
Claude Fable 5: 76.8 (#4), Claude Opus 4.8: 64.6 (#14)
| Benchmark | Claude Fable 5 | Claude Opus 4.8 |
|---|---|---|
| ARC-AGI-2 | 89.2% | 72.1% |
| SimpleBench | 81.9% | 64.8% |
| Kagi LLM Benchmark | 91.4% | 88.8% |
| NYT Connections (extended) | 92.7% | 91.1% |
| ARC-AGI-1 | 98.5% | 92.5% |
| CritPt | 28.6% | 20.9% |
| Chess Puzzles | 41% | 34% |
| EnigmaEval | 39.3% | 23.5% |
| EBR-Bench | 39.5% | 28.6% |
| LMArena Hard Prompts | 1508 | 1481 |
| Mystery Game Puzzles | 52% | 36% |
| DTBench | 98.4% | 94.9% |
| LMCA | 61.1% | 57.5% |
| Surface Evolver Bench | 95% | 87.5% |
| Bench to the Future 3 | 0.13 | 0.14 |
| Epoch Capabilities Index | 163.6 | 158.3 |
| ForecastBench | — | 59.9 |
Math Claude Fable 5 leads
Claude Fable 5: 88.4 (#5), Claude Opus 4.8: 78.2 (#11)
| Benchmark | Claude Fable 5 | Claude Opus 4.8 |
|---|---|---|
| FrontierMath (Tiers 1-3) | 87% | 80% |
| FrontierMath Tier 4 | 90.2% | 56.1% |
| OTIS Mock AIME 2024-2025 | 100% | 98.3% |
| ProofBench | 95% | 69% |
| LMArena Math | 1519 | 1487 |
| MathArena Final-Answer Competitions | — | 91.8% |
| FrontierMath (Feb 2025 set) | — | 47.2% |
| FrontierMath Erdős | 0% | — |
| FrontierMath Tier 4 (v1) | — | 31.3% |
Knowledge Too close to call
Claude Fable 5: 62.3 (#22), Claude Opus 4.8: 61.3 (#26)
| Benchmark | Claude Fable 5 | Claude Opus 4.8 |
|---|---|---|
| GPQA Diamond | 85.9% | 91% |
| SimpleQA Verified | 70.7% | 53% |
| LMArena Expert | 1534 | 1501 |
Multimodal Claude Fable 5 leads
Claude Fable 5: 45.6 (#14), Claude Opus 4.8: 43.2 (#20)
| Benchmark | Claude Fable 5 | Claude Opus 4.8 |
|---|---|---|
| LMArena Vision | 1326 | 1295 |
| Blueprint-Bench 2 | 38.6% | 14.5% |
| Furniture Assembly | 35.8% | 42.5% |
| LMArena Document | 1496 | 1475 |
Multilingual Claude Fable 5 leads
Claude Fable 5: 57.6 (#8), Claude Opus 4.8: 55.4 (#31)
| Benchmark | Claude Fable 5 | Claude Opus 4.8 |
|---|---|---|
| LMArena Non-English | 1483 | 1450 |
| LMArena Chinese | 1546 | 1509 |
| LMArena French | 1509 | 1482 |
| LMArena German | 1486 | 1470 |
| LMArena Japanese | 1507 | 1440 |
| LMArena Korean | 1493 | 1432 |
| LMArena Russian | 1507 | 1476 |
| LMArena Spanish | 1503 | 1470 |
Instruction Following Claude Fable 5 leads
Claude Fable 5: 78.8 (#7), Claude Opus 4.8: 77.6 (#22)
| Benchmark | Claude Fable 5 | Claude Opus 4.8 |
|---|---|---|
| LMArena Instruction Following | 1503 | 1476 |
Long Context Too close to call
Claude Fable 5: 46.3 (#22), Claude Opus 4.8: 45.4 (#37)
| Benchmark | Claude Fable 5 | Claude Opus 4.8 |
|---|---|---|
| LMArena Longer Query | 1510 | 1482 |
Writing & Preference Claude Fable 5 leads
Claude Fable 5: 76.1 (#5), Claude Opus 4.8: 72.1 (#17)
| Benchmark | Claude Fable 5 | Claude Opus 4.8 |
|---|---|---|
| LMArena Text | 1491 | 1460 |
| LMArena Creative Writing | 1494 | 1453 |
| EQ-Bench Creative Writing | 1943 | 1840 |
| EQ-Bench 4 | 1340 | 1281 |
| LMArena Multi-Turn | 1505 | 1476 |
Frequently asked questions
Is Claude Fable 5 better than Claude Opus 4.8?
Claude Fable 5 is the stronger model overall, scoring 66.3 to 60.7 on the Noometry Index. Claude Opus 4.8 costs 2.0× 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 Claude Opus 4.8?
Claude Opus 4.8 is cheaper. It lists at $5 per million input tokens and $25 per million output tokens; Claude Fable 5 lists at $10 and $50.
Is Claude Fable 5 or Claude Opus 4.8 better for coding?
Claude Fable 5 scores higher on coding benchmarks: 67.0 versus 60.2 in the Noometry coding category.
Which has the bigger context window?
Both accept 1M tokens.
How many benchmarks do Claude Fable 5 and Claude Opus 4.8 share?
58 benchmarks have published results for both models. Claude Fable 5 has 61 scored results on Noometry and Claude Opus 4.8 has 65.