Model comparison
Gemini 3.1 Flash Lite vs GPT-6 Astra
GPT-6 Astra is the stronger model overall, scoring 70.8 to 40.8 on the Noometry Index. Gemini 3.1 Flash Lite costs 36× less per token, which makes it the better buy when GPT-6 Astra's lead doesn't matter for your workload.
Last verified . 33 shared benchmarks.
GoogleSummary
- They share 33 benchmarks with published results for both. Gemini 3.1 Flash Lite 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 22.9.
- The biggest single-benchmark swing is NYT Connections (extended): 8.2% for Gemini 3.1 Flash Lite and 98.1% for GPT-6 Astra.
- Gemini 3.1 Flash Lite is cheaper at $0.25 / $1.50 per million input/output tokens, against $10 / $50 for GPT-6 Astra.
- GPT-6 Astra accepts more context: 1.05M tokens versus 1.05M.
Side by side
| Gemini 3.1 Flash Lite | GPT-6 Astra | |
|---|---|---|
| Provider | OpenAI | |
| Noometry Index | 40.8 | 70.8 |
| Released | 2026-03-03 | 2026-09-03 |
| Weights | Proprietary | Proprietary |
| Context window | 1.05M | 1.05M |
| Max output | 66K | 128K |
| Input $ / M tokens | $0.25 | $10 |
| Output $ / M tokens | $1.50 | $50 |
| Results tracked | 38 | 56 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding GPT-6 Astra leads
Gemini 3.1 Flash Lite: 37.8 (#188), GPT-6 Astra: 73.7 (#2)
| Benchmark | Gemini 3.1 Flash Lite | GPT-6 Astra |
|---|---|---|
| LMArena WebDev | 1256 | 1786 |
| SciCode | 41.9% | 56.5% |
| WeirdML | 52.2% | 93.6% |
| LMArena Coding | 1400 | 1487 |
| ALE-Bench | 797.73 | 2,951 |
| DeepSWE | — | 74.1% |
| FrontierCode | — | 53.3% |
| FrontierSWE | — | 65.5% |
| GSO | — | 79.4% |
| MirrorCode | — | 46.7% |
Agentic & Tool Use GPT-6 Astra leads
Gemini 3.1 Flash Lite: 30.2 (#79), GPT-6 Astra: 52.9 (#3)
| Benchmark | Gemini 3.1 Flash Lite | GPT-6 Astra |
|---|---|---|
| APEX-Agents | — | 64.7% |
| Remote Labor Index | — | 20.8% |
| DeepResearch Bench | 37.3% | — |
| BALROG | — | 68.3% |
| GDP.pdf | — | 34.2% |
| Vending-Bench 2 | — | 15,515 |
Reasoning GPT-6 Astra leads
Gemini 3.1 Flash Lite: 22.9 (#186), GPT-6 Astra: 85.1 (#1)
| Benchmark | Gemini 3.1 Flash Lite | GPT-6 Astra |
|---|---|---|
| NYT Connections (extended) | 8.2% | 98.1% |
| CritPt | 1.1% | 31.7% |
| Chess Puzzles | 25% | 72% |
| LMArena Hard Prompts | 1407 | 1462 |
| DTBench | 76.8% | 97.3% |
| LMCA | 35% | 64.4% |
| Epoch Capabilities Index | 144.47 | 166.45 |
| ARC-AGI-2 | — | 95% |
| Kagi LLM Benchmark | 67.2% | — |
| ARC-AGI-1 | — | 98.5% |
| EnigmaEval | 3% | — |
| Thematic Generalization | 63.3% | — |
| EBR-Bench | — | 76.2% |
| Mystery Game Puzzles | — | 84% |
| Bench to the Future 3 | — | 0.14 |
| ForecastBench | 54.4 | — |
Math GPT-6 Astra leads
Gemini 3.1 Flash Lite: 40.7 (#90), GPT-6 Astra: 93.5 (#2)
| Benchmark | Gemini 3.1 Flash Lite | GPT-6 Astra |
|---|---|---|
| FrontierMath (Tiers 1-3) | 27.7% | 93.7% |
| OTIS Mock AIME 2024-2025 | 80% | 100% |
| LMArena Math | 1428 | 1465 |
| FrontierMath Tier 4 | — | 97.6% |
| ProofBench | — | 99% |
| FrontierMath Erdős | — | 2.9% |
Knowledge GPT-6 Astra leads
Gemini 3.1 Flash Lite: 41.9 (#104), GPT-6 Astra: 75.3 (#1)
| Benchmark | Gemini 3.1 Flash Lite | GPT-6 Astra |
|---|---|---|
| GPQA Diamond | 81.8% | 95.8% |
| Humanity's Last Exam | 8.6% | 54.8% |
| Vectara Hallucination Rate | 8.2% | 8.7% |
| LMArena Expert | 1398 | 1483 |
| SimpleQA Verified | — | 75.6% |
Multimodal GPT-6 Astra leads
Gemini 3.1 Flash Lite: 39.4 (#60), GPT-6 Astra: 55.0 (#3)
| Benchmark | Gemini 3.1 Flash Lite | GPT-6 Astra |
|---|---|---|
| LMArena Vision | 1240 | 1281 |
| Blueprint-Bench 2 | — | 49.7% |
| Furniture Assembly | — | 80% |
| LMArena Document | — | 1468 |
Multilingual GPT-6 Astra leads
Gemini 3.1 Flash Lite: 52.3 (#86), GPT-6 Astra: 53.7 (#61)
| Benchmark | Gemini 3.1 Flash Lite | GPT-6 Astra |
|---|---|---|
| LMArena Non-English | 1411 | 1430 |
| LMArena Chinese | 1461 | 1484 |
| LMArena French | 1424 | 1456 |
| LMArena German | 1429 | 1440 |
| LMArena Japanese | 1413 | 1379 |
| LMArena Korean | 1392 | 1426 |
| LMArena Russian | 1420 | 1436 |
| LMArena Spanish | 1421 | 1407 |
Instruction Following GPT-6 Astra leads
Gemini 3.1 Flash Lite: 72.7 (#131), GPT-6 Astra: 76.3 (#44)
| Benchmark | Gemini 3.1 Flash Lite | GPT-6 Astra |
|---|---|---|
| LMArena Instruction Following | 1377 | 1450 |
Long Context GPT-6 Astra leads
Gemini 3.1 Flash Lite: 42.5 (#122), GPT-6 Astra: 44.5 (#62)
| Benchmark | Gemini 3.1 Flash Lite | GPT-6 Astra |
|---|---|---|
| LMArena Longer Query | 1394 | 1456 |
Writing & Preference GPT-6 Astra leads
Gemini 3.1 Flash Lite: 60.9 (#94), GPT-6 Astra: 75.3 (#7)
| Benchmark | Gemini 3.1 Flash Lite | GPT-6 Astra |
|---|---|---|
| LMArena Text | 1416 | 1441 |
| LMArena Creative Writing | 1401 | 1418 |
| LMArena Multi-Turn | 1417 | 1448 |
| EQ-Bench Creative Writing | — | 2173 |
Frequently asked questions
Is Gemini 3.1 Flash Lite better than GPT-6 Astra?
GPT-6 Astra is the stronger model overall, scoring 70.8 to 40.8 on the Noometry Index. Gemini 3.1 Flash Lite costs 36× less per token, which makes it the better buy when GPT-6 Astra's lead doesn't matter for your workload.
Which is cheaper, Gemini 3.1 Flash Lite or GPT-6 Astra?
Gemini 3.1 Flash Lite is cheaper. It lists at $0.25 per million input tokens and $1.50 per million output tokens; GPT-6 Astra lists at $10 and $50.
Is Gemini 3.1 Flash Lite or GPT-6 Astra better for coding?
GPT-6 Astra scores higher on coding benchmarks: 73.7 versus 37.8 in the Noometry coding category.
Which has the bigger context window?
GPT-6 Astra does, with 1.05M tokens against 1.05M.
How many benchmarks do Gemini 3.1 Flash Lite and GPT-6 Astra share?
33 benchmarks have published results for both models. Gemini 3.1 Flash Lite has 38 scored results on Noometry and GPT-6 Astra has 56.