# Claude Fable 5 vs GPT-5.2

> Claude Fable 5 is the stronger model overall, scoring 66.3 to 54.1 on the Noometry Index. GPT-5.2 costs 4.2× less per token, which makes it the better buy when Claude Fable 5's lead doesn't matter for your workload.

- Canonical page: https://noometry.com/compare/claude-fable-5-vs-gpt-5-2
- Last updated: 2026-10-10
- Shared benchmarks: 46

## Summary

- They share 46 benchmarks with published results for both. Claude Fable 5 scores higher in 9 categories and GPT-5.2 in 1 category; 10 gaps are clear of the uncertainty.
- The widest gap is in math, where Claude Fable 5 leads 88.4 to 59.8.
- The biggest single-benchmark swing is ProofBench: 95% for Claude Fable 5 and 15% for GPT-5.2.
- GPT-5.2 is cheaper at $1.75 / $14 per million input/output tokens, against $10 / $50 for Claude Fable 5.
- Claude Fable 5 accepts more context: 1M tokens versus 400K.

## Snapshot

| | Claude Fable 5 | GPT-5.2 |
|---|---|---|
| Provider | Anthropic | OpenAI |
| Noometry Index | 66.3 | 54.1 |
| Rank | 5 | 31 |
| Context | 1M | 400K |
| Input $/M | $10 | $1.75 |
| Output $/M | $50 | $14 |
| Weights | Proprietary | Proprietary |

## Coding

- Claude Fable 5: 67.0 (#4)
- GPT-5.2: 52.2 (#32)

| Benchmark | Claude Fable 5 | GPT-5.2 |
|---|---|---|
| LMArena WebDev | 1626 | 1415 |
| WeirdML | 91.9% | 72.2% |
| LMArena Coding | 1519 | 1447 |
| ALE-Bench | 2,041 | 1,294 |
| SWE-bench Verified | — | 73.8% |
| DeepSWE | 69.9% | — |
| FrontierCode | 53.5% | — |
| SWE-bench Verified (bash only) | — | 72.8% |
| SWE-bench Multilingual | — | 66.7% |
| FrontierSWE | 47% | — |
| SciCode | 61% | — |
| GSO | — | 27.4% |
| MirrorCode | 63.9% | — |
| AlgoTune | — | 2.05 |

## Agentic & Tool Use

- Claude Fable 5: 54.1 (#2)
- GPT-5.2: 40.0 (#24)

| Benchmark | Claude Fable 5 | GPT-5.2 |
|---|---|---|
| Remote Labor Index | 16.1% | 2.5% |
| τ²-bench Banking | 39.7% | 32.2% |
| LMArena Search | 1230 | 1207 |
| Vending-Bench 2 | 5,680 | 3,591 |
| Terminal-Bench | — | 64.9% |
| APEX-Agents | 63.6% | — |
| Berkeley Function Calling Leaderboard | — | 55.9% |
| GDPval | — | 49.7% |
| τ²-bench Airline | — | 83% |
| τ²-bench Retail | — | 81.6% |
| τ²-bench Telecom | — | 89.7% |
| DeepResearch Bench | — | 41.1% |
| PostTrainBench | 41.8% | — |
| GBAEval | 74.5% | — |
| GDP.pdf | 30% | — |
| METR Time Horizons | — | 75.3% |

## Reasoning

- Claude Fable 5: 76.8 (#4)
- GPT-5.2: 50.1 (#33)

| Benchmark | Claude Fable 5 | GPT-5.2 |
|---|---|---|
| ARC-AGI-2 | 89.2% | 52.9% |
| SimpleBench | 81.9% | 45.8% |
| Kagi LLM Benchmark | 91.4% | 73.3% |
| NYT Connections (extended) | 92.7% | 83.6% |
| ARC-AGI-1 | 98.5% | 86.2% |
| Chess Puzzles | 41% | 49% |
| EnigmaEval | 39.3% | 10.4% |
| EBR-Bench | 39.5% | 23% |
| LMArena Hard Prompts | 1508 | 1445 |
| Mystery Game Puzzles | 52% | 23% |
| DTBench | 98.4% | 90.9% |
| LMCA | 61.1% | 43.9% |
| Epoch Capabilities Index | 163.6 | 153.45 |
| CritPt | 28.6% | — |
| Surface Evolver Bench | 95% | — |
| Bench to the Future 3 | 0.13 | — |
| ForecastBench | — | 60.1 |

## Math

- Claude Fable 5: 88.4 (#5)
- GPT-5.2: 59.8 (#32)

| Benchmark | Claude Fable 5 | GPT-5.2 |
|---|---|---|
| FrontierMath (Tiers 1-3) | 87% | 67.4% |
| FrontierMath Tier 4 | 90.2% | 31.7% |
| OTIS Mock AIME 2024-2025 | 100% | 96.1% |
| ProofBench | 95% | 15% |
| LMArena Math | 1519 | 1439 |
| MathArena Final-Answer Competitions | — | 72% |
| FrontierMath (Feb 2025 set) | — | 40.7% |
| FrontierMath Erdős | 0% | — |
| FrontierMath Tier 4 (v1) | — | 18.8% |

## Knowledge

- Claude Fable 5: 62.3 (#22)
- GPT-5.2: 59.2 (#29)

| Benchmark | Claude Fable 5 | GPT-5.2 |
|---|---|---|
| GPQA Diamond | 85.9% | 91.4% |
| SimpleQA Verified | 70.7% | 37.1% |
| LMArena Expert | 1534 | 1445 |
| Humanity's Last Exam | — | 27.8% |
| Vectara Hallucination Rate | — | 8.4% |

## Multimodal

- Claude Fable 5: 45.6 (#14)
- GPT-5.2: 50.9 (#7)

| Benchmark | Claude Fable 5 | GPT-5.2 |
|---|---|---|
| LMArena Vision | 1326 | 1268 |
| Furniture Assembly | 35.8% | 38.3% |
| LMArena Document | 1496 | 1405 |
| VPCT | — | 84% |
| Blueprint-Bench 2 | 38.6% | — |

## Multilingual

- Claude Fable 5: 57.6 (#8)
- GPT-5.2: 53.6 (#63)

| Benchmark | Claude Fable 5 | GPT-5.2 |
|---|---|---|
| LMArena Non-English | 1483 | 1425 |
| LMArena Chinese | 1546 | 1460 |
| LMArena French | 1509 | 1456 |
| LMArena German | 1486 | 1448 |
| LMArena Japanese | 1507 | 1421 |
| LMArena Korean | 1493 | 1390 |
| LMArena Russian | 1507 | 1440 |
| LMArena Spanish | 1503 | 1433 |

## Instruction Following

- Claude Fable 5: 78.8 (#7)
- GPT-5.2: 74.9 (#83)

| Benchmark | Claude Fable 5 | GPT-5.2 |
|---|---|---|
| LMArena Instruction Following | 1503 | 1417 |

## Long Context

- Claude Fable 5: 46.3 (#22)
- GPT-5.2: 44.0 (#74)

| Benchmark | Claude Fable 5 | GPT-5.2 |
|---|---|---|
| LMArena Longer Query | 1510 | 1428 |
| CL-bench | — | 18.2% |

## Writing & Preference

- Claude Fable 5: 76.1 (#5)
- GPT-5.2: 67.0 (#30)

| Benchmark | Claude Fable 5 | GPT-5.2 |
|---|---|---|
| LMArena Text | 1491 | 1439 |
| LMArena Creative Writing | 1494 | 1402 |
| EQ-Bench Creative Writing | 1943 | 1703 |
| LMArena Multi-Turn | 1505 | 1458 |
| EQ-Bench 4 | 1340 | — |

## FAQ

### Is Claude Fable 5 better than GPT-5.2?

Claude Fable 5 is the stronger model overall, scoring 66.3 to 54.1 on the Noometry Index. GPT-5.2 costs 4.2× 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-5.2?

GPT-5.2 is cheaper. It lists at $1.75 per million input tokens and $14 per million output tokens; Claude Fable 5 lists at $10 and $50.

### Is Claude Fable 5 or GPT-5.2 better for coding?

Claude Fable 5 scores higher on coding benchmarks: 67.0 versus 52.2 in the Noometry coding category.

### Which has the bigger context window?

Claude Fable 5 does, with 1M tokens against 400K.

### How many benchmarks do Claude Fable 5 and GPT-5.2 share?

46 benchmarks have published results for both models. Claude Fable 5 has 61 scored results on Noometry and GPT-5.2 has 67.