Model comparison
Gemma 3n E4b IT vs GPT-6 Astra
GPT-6 Astra is the stronger model overall, scoring 70.3 to 36.7 on the Noometry Index. Gemma 3n E4b IT costs 267× less per token, which makes it the better buy when GPT-6 Astra's lead doesn't matter for your workload.
Last verified . 14 shared benchmarks.
GoogleSummary
- They share 14 benchmarks with published results for both. Gemma 3n E4b IT scores higher in 0 categories and GPT-6 Astra in 8 categories; 8 gaps are clear of the uncertainty.
- The widest gap is in reasoning, where GPT-6 Astra leads 85.1 to 19.2.
- Gemma 3n E4b IT is cheaper at $0.06 / $0.12 per million input/output tokens, against $10 / $50 for GPT-6 Astra.
- GPT-6 Astra accepts more context: 1.05M tokens versus 33K.
- Gemma 3n E4b IT has downloadable open weights; the other is API-only.
Side by side
| Gemma 3n E4b IT | GPT-6 Astra | |
|---|---|---|
| Provider | OpenAI | |
| Noometry Index | 36.7 | 70.3 |
| Released | 2025-05-20 | 2026-09-03 |
| Weights | Open | Proprietary |
| Context window | 33K | 1.05M |
| Max output | 33K | 128K |
| Input $ / M tokens | $0.06 | $10 |
| Output $ / M tokens | $0.12 | $50 |
| Results tracked | 18 | 51 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding GPT-6 Astra leads
Gemma 3n E4b IT: 36.4 (#193), GPT-6 Astra: 70.5 (#3)
| Benchmark | Gemma 3n E4b IT | GPT-6 Astra |
|---|---|---|
| LMArena Coding | 1269 | 1488 |
| DeepSWE | — | 74.1% |
| FrontierCode | — | 53.3% |
| LMArena WebDev | — | 1792 |
| FrontierSWE | — | 65.5% |
| SciCode | — | 56.5% |
| WeirdML | — | 93.6% |
| MirrorCode | — | 46.7% |
| ALE-Bench | — | 2,951 |
Agentic & Tool Use Not comparable
Gemma 3n E4b IT: —, GPT-6 Astra: 52.3 (#3)
| Benchmark | Gemma 3n E4b IT | GPT-6 Astra |
|---|---|---|
| APEX-Agents | — | 64.7% |
| Remote Labor Index | — | 20.8% |
| BALROG | — | 68.3% |
| Vending-Bench 2 | — | 15,515 |
Reasoning GPT-6 Astra leads
Gemma 3n E4b IT: 19.2 (#239), GPT-6 Astra: 85.1 (#1)
| Benchmark | Gemma 3n E4b IT | GPT-6 Astra |
|---|---|---|
| LMArena Hard Prompts | 1284 | 1464 |
| ARC-AGI-2 | — | 95% |
| Kagi LLM Benchmark | 31.5% | — |
| NYT Connections (extended) | — | 98.1% |
| ARC-AGI-1 | — | 98.5% |
| CritPt | — | 31.7% |
| Chess Puzzles | — | 72% |
| EBR-Bench | — | 76.2% |
| Mystery Game Puzzles | — | 84% |
| DTBench | — | 97.3% |
| LMCA | — | 64.4% |
| Bench to the Future 3 | — | 0.14 |
| Epoch Capabilities Index | — | 166.6 |
Math GPT-6 Astra leads
Gemma 3n E4b IT: 32.5 (#186), GPT-6 Astra: 93.4 (#1)
| Benchmark | Gemma 3n E4b IT | GPT-6 Astra |
|---|---|---|
| LMArena Math | 1251 | 1463 |
| FrontierMath (Tiers 1-3) | — | 93.7% |
| FrontierMath Tier 4 | — | 97.6% |
| OTIS Mock AIME 2024-2025 | — | 100% |
| ProofBench | — | 99% |
| FrontierMath Erdős | — | 2.9% |
Knowledge GPT-6 Astra leads
Gemma 3n E4b IT: 33.6 (#193), GPT-6 Astra: 75.3 (#1)
| Benchmark | Gemma 3n E4b IT | GPT-6 Astra |
|---|---|---|
| LMArena Expert | 1245 | 1486 |
| GPQA Diamond | — | 95.8% |
| Humanity's Last Exam | — | 54.8% |
| SimpleQA Verified | — | 75.6% |
| Vectara Hallucination Rate | — | 8.7% |
Multimodal Not comparable
Gemma 3n E4b IT: —, GPT-6 Astra: 55.1 (#3)
| Benchmark | Gemma 3n E4b IT | GPT-6 Astra |
|---|---|---|
| LMArena Vision | — | 1280 |
| Blueprint-Bench 2 | — | 49.7% |
| Furniture Assembly | — | 80% |
| LMArena Document | — | 1468 |
Multilingual GPT-6 Astra leads
Gemma 3n E4b IT: 43.5 (#179), GPT-6 Astra: 54.0 (#53)
| Benchmark | Gemma 3n E4b IT | GPT-6 Astra |
|---|---|---|
| LMArena Non-English | 1285 | 1431 |
| LMArena Chinese | 1310 | 1495 |
| LMArena French | 1330 | 1467 |
| LMArena Russian | 1288 | 1434 |
| LMArena Spanish | 1305 | 1412 |
| LMArena German | 1312 | — |
| LMArena Japanese | 1274 | — |
| LMArena Korean | 1259 | — |
Instruction Following GPT-6 Astra leads
Gemma 3n E4b IT: 66.3 (#205), GPT-6 Astra: 76.6 (#37)
| Benchmark | Gemma 3n E4b IT | GPT-6 Astra |
|---|---|---|
| LMArena Instruction Following | 1255 | 1454 |
Long Context GPT-6 Astra leads
Gemma 3n E4b IT: 38.8 (#185), GPT-6 Astra: 44.7 (#57)
| Benchmark | Gemma 3n E4b IT | GPT-6 Astra |
|---|---|---|
| LMArena Longer Query | 1277 | 1460 |
Writing & Preference GPT-6 Astra leads
Gemma 3n E4b IT: 50.3 (#181), GPT-6 Astra: 75.7 (#7)
| Benchmark | Gemma 3n E4b IT | GPT-6 Astra |
|---|---|---|
| LMArena Text | 1306 | 1443 |
| LMArena Creative Writing | 1287 | 1423 |
| LMArena Multi-Turn | 1276 | 1457 |
| EQ-Bench Creative Writing | — | 2173 |
Frequently asked questions
Is Gemma 3n E4b IT better than GPT-6 Astra?
GPT-6 Astra is the stronger model overall, scoring 70.3 to 36.7 on the Noometry Index. Gemma 3n E4b IT costs 267× less per token, which makes it the better buy when GPT-6 Astra's lead doesn't matter for your workload.
Which is cheaper, Gemma 3n E4b IT or GPT-6 Astra?
Gemma 3n E4b IT is cheaper. It lists at $0.06 per million input tokens and $0.12 per million output tokens; GPT-6 Astra lists at $10 and $50.
Is Gemma 3n E4b IT or GPT-6 Astra better for coding?
GPT-6 Astra scores higher on coding benchmarks: 70.5 versus 36.4 in the Noometry coding category.
Which has the bigger context window?
GPT-6 Astra does, with 1.05M tokens against 33K.
How many benchmarks do Gemma 3n E4b IT and GPT-6 Astra share?
14 benchmarks have published results for both models. Gemma 3n E4b IT has 18 scored results on Noometry and GPT-6 Astra has 51.