# GPT-6 Astra vs phi-3-medium 14B

> GPT-6 Astra is the stronger model overall, scoring 70.8 to 29.7 on the Noometry Index.

- Canonical page: https://noometry.com/compare/gpt-6-astra-vs-phi-3-medium-14b
- Last updated: 2026-10-10
- Shared benchmarks: 2

## Summary

- They share 2 benchmarks with published results for both. GPT-6 Astra scores higher in 3 categories and phi-3-medium 14B in 0 categories; 3 gaps are clear of the uncertainty.
- The widest gap is in knowledge, where GPT-6 Astra leads 75.3 to 9.1.
- The biggest single-benchmark swing is GPQA Diamond: 95.8% for GPT-6 Astra and 27.6% for phi-3-medium 14B.
- phi-3-medium 14B has downloadable open weights; the other is API-only.

## Snapshot

| | GPT-6 Astra | phi-3-medium 14B |
|---|---|---|
| Provider | OpenAI | Microsoft |
| Noometry Index | 70.8 | 29.7 |
| Rank | 1 | 306 |
| Context | 1.05M | — |
| Input $/M | $10 | — |
| Output $/M | $50 | — |
| Weights | Proprietary | Open |

## Coding

- GPT-6 Astra: 73.7 (#2)
- phi-3-medium 14B: 36.8 (#201)

| Benchmark | GPT-6 Astra | phi-3-medium 14B |
|---|---|---|
| DeepSWE | 74.1% | — |
| FrontierCode | 53.3% | — |
| LMArena WebDev | 1786 | — |
| FrontierSWE | 65.5% | — |
| SciCode | 56.5% | — |
| GSO | 79.4% | — |
| WeirdML | 93.6% | — |
| BigCodeBench Instruct | — | 37.6% |
| LMArena Coding | 1487 | — |
| MirrorCode | 46.7% | — |
| BigCodeBench Complete | — | 48.7% |
| ALE-Bench | 2,951 | — |

## Agentic & Tool Use

- GPT-6 Astra: 52.9 (#3)
- phi-3-medium 14B: —

| Benchmark | GPT-6 Astra | phi-3-medium 14B |
|---|---|---|
| APEX-Agents | 64.7% | — |
| Remote Labor Index | 20.8% | — |
| BALROG | 68.3% | — |
| GDP.pdf | 34.2% | — |
| Vending-Bench 2 | 15,515 | — |

## Reasoning

- GPT-6 Astra: 85.1 (#1)
- phi-3-medium 14B: —

| Benchmark | GPT-6 Astra | phi-3-medium 14B |
|---|---|---|
| Epoch Capabilities Index | 166.45 | 121.23 |
| ARC-AGI-2 | 95% | — |
| NYT Connections (extended) | 98.1% | — |
| ARC-AGI-1 | 98.5% | — |
| CritPt | 31.7% | — |
| Chess Puzzles | 72% | — |
| EBR-Bench | 76.2% | — |
| LMArena Hard Prompts | 1462 | — |
| Mystery Game Puzzles | 84% | — |
| DTBench | 97.3% | — |
| LMCA | 64.4% | — |
| Adversarial NLI | — | 55.8% |
| Bench to the Future 3 | 0.14 | — |
| BIG-Bench Hard | — | 81.4% |
| HellaSwag | — | 82.4% |
| WinoGrande | — | 81.5% |

## Math

- GPT-6 Astra: 93.5 (#2)
- phi-3-medium 14B: 27.3 (#250)

| Benchmark | GPT-6 Astra | phi-3-medium 14B |
|---|---|---|
| FrontierMath (Tiers 1-3) | 93.7% | — |
| FrontierMath Tier 4 | 97.6% | — |
| OTIS Mock AIME 2024-2025 | 100% | — |
| ProofBench | 99% | — |
| LMArena Math | 1465 | — |
| MATH Level 5 | — | 17.6% |
| FrontierMath Erdős | 2.9% | — |

## Knowledge

- GPT-6 Astra: 75.3 (#1)
- phi-3-medium 14B: 9.1 (#306)

| Benchmark | GPT-6 Astra | phi-3-medium 14B |
|---|---|---|
| GPQA Diamond | 95.8% | 27.6% |
| Humanity's Last Exam | 54.8% | — |
| SimpleQA Verified | 75.6% | — |
| Vectara Hallucination Rate | 8.7% | — |
| LMArena Expert | 1483 | — |
| ARC (AI2) Challenge | — | 91.6% |
| MMLU | — | 78% |
| OpenBookQA | — | 87.4% |
| TriviaQA | — | 73.9% |

## Multimodal

- GPT-6 Astra: 55.0 (#3)
- phi-3-medium 14B: —

| Benchmark | GPT-6 Astra | phi-3-medium 14B |
|---|---|---|
| LMArena Vision | 1281 | — |
| Blueprint-Bench 2 | 49.7% | — |
| Furniture Assembly | 80% | — |
| LMArena Document | 1468 | — |

## Multilingual

- GPT-6 Astra: 53.7 (#61)
- phi-3-medium 14B: —

| Benchmark | GPT-6 Astra | phi-3-medium 14B |
|---|---|---|
| LMArena Non-English | 1430 | — |
| LMArena Chinese | 1484 | — |
| LMArena French | 1456 | — |
| LMArena German | 1440 | — |
| LMArena Japanese | 1379 | — |
| LMArena Korean | 1426 | — |
| LMArena Russian | 1436 | — |
| LMArena Spanish | 1407 | — |

## Instruction Following

- GPT-6 Astra: 76.3 (#44)
- phi-3-medium 14B: —

| Benchmark | GPT-6 Astra | phi-3-medium 14B |
|---|---|---|
| LMArena Instruction Following | 1450 | — |

## Long Context

- GPT-6 Astra: 44.5 (#62)
- phi-3-medium 14B: —

| Benchmark | GPT-6 Astra | phi-3-medium 14B |
|---|---|---|
| LMArena Longer Query | 1456 | — |

## Writing & Preference

- GPT-6 Astra: 75.3 (#7)
- phi-3-medium 14B: —

| Benchmark | GPT-6 Astra | phi-3-medium 14B |
|---|---|---|
| LMArena Text | 1441 | — |
| LMArena Creative Writing | 1418 | — |
| EQ-Bench Creative Writing | 2173 | — |
| LMArena Multi-Turn | 1448 | — |

## FAQ

### Is GPT-6 Astra better than phi-3-medium 14B?

GPT-6 Astra is the stronger model overall, scoring 70.8 to 29.7 on the Noometry Index.

### Is GPT-6 Astra or phi-3-medium 14B better for coding?

GPT-6 Astra scores higher on coding benchmarks: 73.7 versus 36.8 in the Noometry coding category.

### How many benchmarks do GPT-6 Astra and phi-3-medium 14B share?

2 benchmarks have published results for both models. GPT-6 Astra has 56 scored results on Noometry and phi-3-medium 14B has 13.