# Claude Fable 5 vs gpt-oss-20b

> Claude Fable 5 is the stronger model overall, scoring 66.8 to 32.5 on the Noometry Index. gpt-oss-20b costs 556× 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-oss-20b
- Last updated: 2026-10-12
- Shared benchmarks: 28

## Summary

- They share 28 benchmarks with published results for both. Claude Fable 5 scores higher in 9 categories and gpt-oss-20b in 0 categories; 9 gaps are clear of the uncertainty.
- The widest gap is in reasoning, where Claude Fable 5 leads 76.8 to 19.3.
- The biggest single-benchmark swing is WeirdML: 91.9% for Claude Fable 5 and 40.9% for gpt-oss-20b.
- gpt-oss-20b is cheaper at $0.018 / $0.09 per million input/output tokens, against $10 / $50 for Claude Fable 5.
- Claude Fable 5 accepts more context: 1M tokens versus 131K.
- gpt-oss-20b has downloadable open weights; the other is API-only.

## Snapshot

| | Claude Fable 5 | gpt-oss-20b |
|---|---|---|
| Provider | Anthropic | OpenAI |
| Noometry Index | 66.8 | 32.5 |
| Rank | 5 | 255 |
| Context | 1M | 131K |
| Input $/M | $10 | $0.018 |
| Output $/M | $50 | $0.09 |
| Weights | Proprietary | Open |

## Coding

- Claude Fable 5: 70.6 (#4)
- gpt-oss-20b: 37.6 (#192)

| Benchmark | Claude Fable 5 | gpt-oss-20b |
|---|---|---|
| SciCode | 61% | 34.4% |
| WeirdML | 91.9% | 40.9% |
| LMArena Coding | 1519 | 1306 |
| ALE-Bench | 2,041 | 566.05 |
| DeepSWE | 69.9% | — |
| FrontierCode | 53.5% | — |
| LMArena WebDev | 1625 | — |
| FrontierSWE | 47% | — |
| GSO | 78.4% | — |
| MirrorCode | 63.9% | — |

## Agentic & Tool Use

- Claude Fable 5: 54.0 (#2)
- gpt-oss-20b: 9.3 (#154)

| Benchmark | Claude Fable 5 | gpt-oss-20b |
|---|---|---|
| Terminal-Bench | — | 3.4% |
| APEX-Agents | 63.6% | — |
| Remote Labor Index | 16.1% | — |
| τ²-bench Banking | 39.7% | — |
| PostTrainBench | 41.8% | — |
| GBAEval | 74.5% | — |
| GDP.pdf | 30% | — |
| LMArena Search | 1230 | — |
| Vending-Bench 2 | 5,680 | — |

## Reasoning

- Claude Fable 5: 76.8 (#6)
- gpt-oss-20b: 19.3 (#261)

| Benchmark | Claude Fable 5 | gpt-oss-20b |
|---|---|---|
| Kagi LLM Benchmark | 91.4% | 53.2% |
| CritPt | 28.6% | 1.4% |
| Chess Puzzles | 41% | 4% |
| LMArena Hard Prompts | 1508 | 1274 |
| DTBench | 98.4% | 68% |
| LMCA | 61.1% | 14.5% |
| Epoch Capabilities Index | 162.06 | 137.82 |
| ARC-AGI-2 | 89.2% | — |
| SimpleBench | 81.9% | — |
| NYT Connections (extended) | 92.7% | — |
| ARC-AGI-1 | 98.5% | — |
| EnigmaEval | 39.3% | — |
| EBR-Bench | 39.5% | — |
| Mystery Game Puzzles | 52% | — |
| Surface Evolver Bench | 95% | — |
| Bench to the Future 3 | 0.13 | — |

## Math

- Claude Fable 5: 88.5 (#5)
- gpt-oss-20b: 39.4 (#103)

| Benchmark | Claude Fable 5 | gpt-oss-20b |
|---|---|---|
| OTIS Mock AIME 2024-2025 | 100% | 65.3% |
| LMArena Math | 1519 | 1317 |
| FrontierMath (Tiers 1-3) | 87% | — |
| FrontierMath Tier 4 | 90.2% | — |
| ProofBench | 95% | — |
| Omni-MATH | — | 56.5% |
| FrontierMath Erdős | 0% | — |

## Knowledge

- Claude Fable 5: 62.2 (#25)
- gpt-oss-20b: 34.6 (#195)

| Benchmark | Claude Fable 5 | gpt-oss-20b |
|---|---|---|
| GPQA Diamond | 85.9% | 60.8% |
| LMArena Expert | 1534 | 1258 |
| SimpleQA Verified | 70.7% | — |
| MMLU-Pro | — | 74% |
| GPQA (HELM) | — | 59.4% |

## Multimodal

- Claude Fable 5: 45.3 (#17)
- gpt-oss-20b: —

| Benchmark | Claude Fable 5 | gpt-oss-20b |
|---|---|---|
| LMArena Vision | 1324 | — |
| Blueprint-Bench 2 | 38.6% | — |
| Furniture Assembly | 35.8% | — |
| LMArena Document | 1496 | — |

## Multilingual

- Claude Fable 5: 57.3 (#9)
- gpt-oss-20b: 42.2 (#197)

| Benchmark | Claude Fable 5 | gpt-oss-20b |
|---|---|---|
| LMArena Non-English | 1481 | 1268 |
| LMArena Chinese | 1543 | 1314 |
| LMArena German | 1486 | 1255 |
| LMArena Japanese | 1506 | 1244 |
| LMArena Korean | 1488 | 1236 |
| LMArena Russian | 1504 | 1278 |
| LMArena Spanish | 1498 | 1267 |
| LMArena French | 1505 | — |

## Instruction Following

- Claude Fable 5: 78.6 (#8)
- gpt-oss-20b: 61.8 (#240)

| Benchmark | Claude Fable 5 | gpt-oss-20b |
|---|---|---|
| LMArena Instruction Following | 1502 | 1236 |
| IFEval | — | 73.2% |

## Long Context

- Claude Fable 5: 46.3 (#23)
- gpt-oss-20b: 37.9 (#209)

| Benchmark | Claude Fable 5 | gpt-oss-20b |
|---|---|---|
| LMArena Longer Query | 1509 | 1250 |

## Writing & Preference

- Claude Fable 5: 75.9 (#5)
- gpt-oss-20b: 35.5 (#265)

| Benchmark | Claude Fable 5 | gpt-oss-20b |
|---|---|---|
| LMArena Text | 1491 | 1287 |
| LMArena Creative Writing | 1494 | 1201 |
| EQ-Bench Creative Writing | 1943 | 666 |
| LMArena Multi-Turn | 1504 | 1268 |
| WildBench | — | 73.7% |
| EQ-Bench 4 | 1340 | — |

## FAQ

### Is Claude Fable 5 better than gpt-oss-20b?

Claude Fable 5 is the stronger model overall, scoring 66.8 to 32.5 on the Noometry Index. gpt-oss-20b costs 556× 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-oss-20b?

gpt-oss-20b is cheaper. It lists at $0.018 per million input tokens and $0.09 per million output tokens; Claude Fable 5 lists at $10 and $50.

### Is Claude Fable 5 or gpt-oss-20b better for coding?

Claude Fable 5 scores higher on coding benchmarks: 70.6 versus 37.6 in the Noometry coding category.

### Which has the bigger context window?

Claude Fable 5 does, with 1M tokens against 131K.

### How many benchmarks do Claude Fable 5 and gpt-oss-20b share?

28 benchmarks have published results for both models. Claude Fable 5 has 62 scored results on Noometry and gpt-oss-20b has 34.