Model comparison
GPT-5 Nano vs GPT-6 Astra
GPT-6 Astra is the stronger model overall, scoring 70.8 to 33.5 on the Noometry Index. GPT-5 Nano costs 145× less per token, which makes it the better buy when GPT-6 Astra's lead doesn't matter for your workload.
Last verified . 34 shared benchmarks.
Summary
- They share 34 benchmarks with published results for both. GPT-5 Nano scores higher in 0 categories and GPT-6 Astra in 10 categories; 10 gaps are clear of the uncertainty.
- The widest gap is in reasoning, where GPT-6 Astra leads 85.1 to 16.3.
- The biggest single-benchmark swing is FrontierMath Tier 4: 2.4% for GPT-5 Nano and 97.6% for GPT-6 Astra.
- GPT-5 Nano is cheaper at $0.05 / $0.40 per million input/output tokens, against $10 / $50 for GPT-6 Astra.
- GPT-6 Astra accepts more context: 1.05M tokens versus 400K.
Side by side
| GPT-5 Nano | GPT-6 Astra | |
|---|---|---|
| Provider | OpenAI | OpenAI |
| Noometry Index | 33.5 | 70.8 |
| Released | 2025-08-07 | 2026-09-03 |
| Weights | Proprietary | Proprietary |
| Context window | 400K | 1.05M |
| Max output | 128K | 128K |
| Input $ / M tokens | $0.05 | $10 |
| Output $ / M tokens | $0.40 | $50 |
| Results tracked | 49 | 56 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding GPT-6 Astra leads
GPT-5 Nano: 33.6 (#254), GPT-6 Astra: 73.7 (#2)
| Benchmark | GPT-5 Nano | GPT-6 Astra |
|---|---|---|
| WeirdML | 38.1% | 93.6% |
| LMArena Coding | 1351 | 1487 |
| ALE-Bench | 718.67 | 2,951 |
| DeepSWE | — | 74.1% |
| FrontierCode | — | 53.3% |
| SWE-bench Verified (bash only) | 34.8% | — |
| LMArena WebDev | — | 1786 |
| FrontierSWE | — | 65.5% |
| SciCode | — | 56.5% |
| GSO | — | 79.4% |
| MirrorCode | — | 46.7% |
Agentic & Tool Use GPT-6 Astra leads
GPT-5 Nano: 25.8 (#106), GPT-6 Astra: 52.9 (#3)
| Benchmark | GPT-5 Nano | GPT-6 Astra |
|---|---|---|
| Terminal-Bench | 21.8% | — |
| APEX-Agents | — | 64.7% |
| Berkeley Function Calling Leaderboard | 51.5% | — |
| Remote Labor Index | — | 20.8% |
| BALROG | — | 68.3% |
| GDP.pdf | — | 34.2% |
| Vending-Bench 2 | — | 15,515 |
Reasoning GPT-6 Astra leads
GPT-5 Nano: 16.3 (#306), GPT-6 Astra: 85.1 (#1)
| Benchmark | GPT-5 Nano | GPT-6 Astra |
|---|---|---|
| ARC-AGI-2 | 2.6% | 95% |
| ARC-AGI-1 | 20.7% | 98.5% |
| Chess Puzzles | 27% | 72% |
| LMArena Hard Prompts | 1328 | 1462 |
| Mystery Game Puzzles | 9% | 84% |
| DTBench | 62.7% | 97.3% |
| LMCA | 7.9% | 64.4% |
| Epoch Capabilities Index | 139.38 | 166.45 |
| Kagi LLM Benchmark | 62.2% | — |
| NYT Connections (extended) | — | 98.1% |
| CritPt | — | 31.7% |
| EBR-Bench | — | 76.2% |
| Bench to the Future 3 | — | 0.14 |
| ForecastBench | 59.1 | — |
Math GPT-6 Astra leads
GPT-5 Nano: 29.4 (#241), GPT-6 Astra: 93.5 (#2)
| Benchmark | GPT-5 Nano | GPT-6 Astra |
|---|---|---|
| FrontierMath (Tiers 1-3) | 20% | 93.7% |
| FrontierMath Tier 4 | 2.4% | 97.6% |
| OTIS Mock AIME 2024-2025 | 81.1% | 100% |
| ProofBench | 12% | 99% |
| LMArena Math | 1317 | 1465 |
| Omni-MATH | 54.6% | — |
| MATH Level 5 | 95.2% | — |
| FrontierMath (Feb 2025 set) | 8.3% | — |
| FrontierMath Erdős | — | 2.9% |
| FrontierMath Tier 4 (v1) | 2.1% | — |
Knowledge GPT-6 Astra leads
GPT-5 Nano: 35.9 (#178), GPT-6 Astra: 75.3 (#1)
| Benchmark | GPT-5 Nano | GPT-6 Astra |
|---|---|---|
| GPQA Diamond | 69.4% | 95.8% |
| SimpleQA Verified | 11.7% | 75.6% |
| Vectara Hallucination Rate | 10.5% | 8.7% |
| LMArena Expert | 1321 | 1483 |
| Humanity's Last Exam | — | 54.8% |
| MMLU-Pro | 77.8% | — |
| GPQA (HELM) | 67.9% | — |
Multimodal GPT-6 Astra leads
GPT-5 Nano: 31.3 (#108), GPT-6 Astra: 55.0 (#3)
| Benchmark | GPT-5 Nano | GPT-6 Astra |
|---|---|---|
| LMArena Vision | 1159 | 1281 |
| VPCT | 37.2% | — |
| Blueprint-Bench 2 | — | 49.7% |
| Furniture Assembly | — | 80% |
| LMArena Document | — | 1468 |
Multilingual GPT-6 Astra leads
GPT-5 Nano: 45.3 (#172), GPT-6 Astra: 53.7 (#61)
| Benchmark | GPT-5 Nano | GPT-6 Astra |
|---|---|---|
| LMArena Non-English | 1313 | 1430 |
| LMArena Chinese | 1356 | 1484 |
| LMArena German | 1327 | 1440 |
| LMArena Japanese | 1226 | 1379 |
| LMArena Korean | 1269 | 1426 |
| LMArena Russian | 1296 | 1436 |
| LMArena Spanish | 1360 | 1407 |
| LMArena French | — | 1456 |
Instruction Following GPT-6 Astra leads
GPT-5 Nano: 75.0 (#79), GPT-6 Astra: 76.3 (#44)
| Benchmark | GPT-5 Nano | GPT-6 Astra |
|---|---|---|
| LMArena Instruction Following | 1306 | 1450 |
| IFEval | 93.2% | — |
Long Context GPT-6 Astra leads
GPT-5 Nano: 31.3 (#281), GPT-6 Astra: 44.5 (#62)
| Benchmark | GPT-5 Nano | GPT-6 Astra |
|---|---|---|
| LMArena Longer Query | 1312 | 1456 |
| Fiction.LiveBench | 44.4% | — |
Writing & Preference GPT-6 Astra leads
GPT-5 Nano: 39.1 (#249), GPT-6 Astra: 75.3 (#7)
| Benchmark | GPT-5 Nano | GPT-6 Astra |
|---|---|---|
| LMArena Text | 1320 | 1441 |
| LMArena Creative Writing | 1249 | 1418 |
| EQ-Bench Creative Writing | 705 | 2173 |
| LMArena Multi-Turn | 1311 | 1448 |
| WildBench | 80.6% | — |
Frequently asked questions
Is GPT-5 Nano better than GPT-6 Astra?
GPT-6 Astra is the stronger model overall, scoring 70.8 to 33.5 on the Noometry Index. GPT-5 Nano costs 145× less per token, which makes it the better buy when GPT-6 Astra's lead doesn't matter for your workload.
Which is cheaper, GPT-5 Nano or GPT-6 Astra?
GPT-5 Nano is cheaper. It lists at $0.05 per million input tokens and $0.40 per million output tokens; GPT-6 Astra lists at $10 and $50.
Is GPT-5 Nano or GPT-6 Astra better for coding?
GPT-6 Astra scores higher on coding benchmarks: 73.7 versus 33.6 in the Noometry coding category.
Which has the bigger context window?
GPT-6 Astra does, with 1.05M tokens against 400K.
How many benchmarks do GPT-5 Nano and GPT-6 Astra share?
34 benchmarks have published results for both models. GPT-5 Nano has 49 scored results on Noometry and GPT-6 Astra has 56.