Model comparison
Gemini 3.1 Flash Lite vs GPT-5.4 nano
GPT-5.4 nano is the stronger model overall, scoring 41.9 to 40.8 on the Noometry Index.
Last verified . 32 shared benchmarks.
GoogleSummary
- They share 32 benchmarks with published results for both. Gemini 3.1 Flash Lite scores higher in 5 categories and GPT-5.4 nano in 4 categories; 4 gaps are clear of the uncertainty.
- The widest gap is in coding, where GPT-5.4 nano leads 43.6 to 37.8.
- The biggest single-benchmark swing is Kagi LLM Benchmark: 67.2% for Gemini 3.1 Flash Lite and 39.7% for GPT-5.4 nano.
- GPT-5.4 nano is cheaper at $0.20 / $1.25 per million input/output tokens, against $0.25 / $1.50 for Gemini 3.1 Flash Lite.
- Gemini 3.1 Flash Lite accepts more context: 1.05M tokens versus 400K.
Side by side
| Gemini 3.1 Flash Lite | GPT-5.4 nano | |
|---|---|---|
| Provider | OpenAI | |
| Noometry Index | 40.8 | 41.9 |
| Released | 2026-03-03 | 2026-03-17 |
| Weights | Proprietary | Proprietary |
| Context window | 1.05M | 400K |
| Max output | 66K | 128K |
| Input $ / M tokens | $0.25 | $0.20 |
| Output $ / M tokens | $1.50 | $1.25 |
| Results tracked | 38 | 40 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding GPT-5.4 nano leads
Gemini 3.1 Flash Lite: 37.8 (#188), GPT-5.4 nano: 43.6 (#84)
| Benchmark | Gemini 3.1 Flash Lite | GPT-5.4 nano |
|---|---|---|
| SciCode | 41.9% | 46.9% |
| WeirdML | 52.2% | 49.2% |
| LMArena Coding | 1400 | 1405 |
| ALE-Bench | 797.73 | 1,005 |
| LMArena WebDev | 1256 | — |
Agentic & Tool Use Not comparable
Gemini 3.1 Flash Lite: 30.2 (#79), GPT-5.4 nano: —
| Benchmark | Gemini 3.1 Flash Lite | GPT-5.4 nano |
|---|---|---|
| DeepResearch Bench | 37.3% | — |
Reasoning Too close to call
Gemini 3.1 Flash Lite: 22.9 (#186), GPT-5.4 nano: 23.7 (#173)
| Benchmark | Gemini 3.1 Flash Lite | GPT-5.4 nano |
|---|---|---|
| Kagi LLM Benchmark | 67.2% | 39.7% |
| CritPt | 1.1% | 9.3% |
| Chess Puzzles | 25% | 30% |
| LMArena Hard Prompts | 1407 | 1381 |
| DTBench | 76.8% | 80.3% |
| LMCA | 35% | 36.9% |
| Epoch Capabilities Index | 144.47 | 145.81 |
| ForecastBench | 54.4 | 57.3 |
| ARC-AGI-2 | — | 5.7% |
| NYT Connections (extended) | 8.2% | — |
| ARC-AGI-1 | — | 51.5% |
| EnigmaEval | 3% | — |
| Thematic Generalization | 63.3% | — |
| Mystery Game Puzzles | — | 9% |
Math Too close to call
Gemini 3.1 Flash Lite: 40.7 (#90), GPT-5.4 nano: 40.9 (#88)
| Benchmark | Gemini 3.1 Flash Lite | GPT-5.4 nano |
|---|---|---|
| FrontierMath (Tiers 1-3) | 27.7% | 44.9% |
| OTIS Mock AIME 2024-2025 | 80% | 87.8% |
| LMArena Math | 1428 | 1406 |
| FrontierMath Tier 4 | — | 12.2% |
| ProofBench | — | 5% |
| FrontierMath (Feb 2025 set) | — | 25.9% |
| FrontierMath Tier 4 (v1) | — | 6.3% |
Knowledge Too close to call
Gemini 3.1 Flash Lite: 41.9 (#104), GPT-5.4 nano: 41.9 (#103)
| Benchmark | Gemini 3.1 Flash Lite | GPT-5.4 nano |
|---|---|---|
| GPQA Diamond | 81.8% | 78.5% |
| Vectara Hallucination Rate | 8.2% | 3.1% |
| LMArena Expert | 1398 | 1396 |
| Humanity's Last Exam | 8.6% | — |
| SimpleQA Verified | — | 11.7% |
Multimodal Gemini 3.1 Flash Lite leads
Gemini 3.1 Flash Lite: 39.4 (#60), GPT-5.4 nano: 36.7 (#78)
| Benchmark | Gemini 3.1 Flash Lite | GPT-5.4 nano |
|---|---|---|
| LMArena Vision | 1240 | 1196 |
Multilingual Gemini 3.1 Flash Lite leads
Gemini 3.1 Flash Lite: 52.3 (#86), GPT-5.4 nano: 48.6 (#140)
| Benchmark | Gemini 3.1 Flash Lite | GPT-5.4 nano |
|---|---|---|
| LMArena Non-English | 1411 | 1359 |
| LMArena Chinese | 1461 | 1392 |
| LMArena French | 1424 | 1396 |
| LMArena German | 1429 | 1367 |
| LMArena Japanese | 1413 | 1343 |
| LMArena Korean | 1392 | 1320 |
| LMArena Russian | 1420 | 1363 |
| LMArena Spanish | 1421 | 1371 |
Instruction Following Too close to call
Gemini 3.1 Flash Lite: 72.7 (#131), GPT-5.4 nano: 71.9 (#144)
| Benchmark | Gemini 3.1 Flash Lite | GPT-5.4 nano |
|---|---|---|
| LMArena Instruction Following | 1377 | 1362 |
Long Context Too close to call
Gemini 3.1 Flash Lite: 42.5 (#122), GPT-5.4 nano: 41.6 (#137)
| Benchmark | Gemini 3.1 Flash Lite | GPT-5.4 nano |
|---|---|---|
| LMArena Longer Query | 1394 | 1366 |
Writing & Preference Gemini 3.1 Flash Lite leads
Gemini 3.1 Flash Lite: 60.9 (#94), GPT-5.4 nano: 55.7 (#142)
| Benchmark | Gemini 3.1 Flash Lite | GPT-5.4 nano |
|---|---|---|
| LMArena Text | 1416 | 1372 |
| LMArena Creative Writing | 1401 | 1314 |
| LMArena Multi-Turn | 1417 | 1382 |
Frequently asked questions
Is Gemini 3.1 Flash Lite better than GPT-5.4 nano?
GPT-5.4 nano is the stronger model overall, scoring 41.9 to 40.8 on the Noometry Index.
Which is cheaper, Gemini 3.1 Flash Lite or GPT-5.4 nano?
GPT-5.4 nano is cheaper. It lists at $0.20 per million input tokens and $1.25 per million output tokens; Gemini 3.1 Flash Lite lists at $0.25 and $1.50.
Is Gemini 3.1 Flash Lite or GPT-5.4 nano better for coding?
GPT-5.4 nano scores higher on coding benchmarks: 43.6 versus 37.8 in the Noometry coding category.
Which has the bigger context window?
Gemini 3.1 Flash Lite does, with 1.05M tokens against 400K.
How many benchmarks do Gemini 3.1 Flash Lite and GPT-5.4 nano share?
32 benchmarks have published results for both models. Gemini 3.1 Flash Lite has 38 scored results on Noometry and GPT-5.4 nano has 40.