Model comparison
Gemini 3.6 Flash vs GPT-5.4 Pro
GPT-5.4 Pro is the stronger model overall, scoring 58.9 to 54.1 on the Noometry Index. Gemini 3.6 Flash costs 45× less per token, which makes it the better buy when GPT-5.4 Pro's lead doesn't matter for your workload.
Last verified . 10 shared benchmarks.
GoogleSummary
- They share 10 benchmarks with published results for both. Gemini 3.6 Flash scores higher in 1 category and GPT-5.4 Pro in 3 categories; 3 gaps are clear of the uncertainty.
- The widest gap is in math, where GPT-5.4 Pro leads 72.4 to 57.3.
- The biggest single-benchmark swing is FrontierMath Tier 4: 22% for Gemini 3.6 Flash and 58.5% for GPT-5.4 Pro.
- Gemini 3.6 Flash is cheaper at $0.75 / $3.75 per million input/output tokens, against $30 / $180 for GPT-5.4 Pro.
- GPT-5.4 Pro accepts more context: 1.05M tokens versus 1.05M.
Side by side
| Gemini 3.6 Flash | GPT-5.4 Pro | |
|---|---|---|
| Provider | OpenAI | |
| Noometry Index | 54.1 | 58.9 |
| Released | 2026-07-21 | 2026-03-05 |
| Weights | Proprietary | Proprietary |
| Context window | 1.05M | 1.05M |
| Max output | 66K | 128K |
| Input $ / M tokens | $0.75 | $30 |
| Output $ / M tokens | $3.75 | $180 |
| Results tracked | 46 | 16 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding Gemini 3.6 Flash leads
Gemini 3.6 Flash: 50.0 (#48), GPT-5.4 Pro: 43.2 (#90)
| Benchmark | Gemini 3.6 Flash | GPT-5.4 Pro |
|---|---|---|
| WeirdML | 56.1% | 57.4% |
| DeepSWE | 46.7% | — |
| FrontierCode | 34.4% | — |
| LMArena WebDev | 1538 | — |
| SciCode | 52.7% | — |
| LMArena Coding | 1491 | — |
| ALE-Bench | 715.52 | — |
Agentic & Tool Use Not comparable
Gemini 3.6 Flash: 32.3 (#65), GPT-5.4 Pro: —
| Benchmark | Gemini 3.6 Flash | GPT-5.4 Pro |
|---|---|---|
| APEX-Agents | 46.9% | — |
| GDP.pdf | 14% | — |
Reasoning GPT-5.4 Pro leads
Gemini 3.6 Flash: 58.8 (#22), GPT-5.4 Pro: 70.7 (#13)
| Benchmark | Gemini 3.6 Flash | GPT-5.4 Pro |
|---|---|---|
| ARC-AGI-2 | 60.4% | 83.3% |
| ARC-AGI-1 | 91.2% | 94.5% |
| CritPt | 10.6% | 30% |
| Chess Puzzles | 43% | 58.6% |
| Epoch Capabilities Index | 154.25 | 158.93 |
| SimpleBench | — | 74.1% |
| NYT Connections (extended) | 89% | — |
| EnigmaEval | — | 23.8% |
| LMArena Hard Prompts | 1485 | — |
| Mystery Game Puzzles | 30% | — |
| DTBench | 95.5% | — |
| LMCA | 44.9% | — |
Math GPT-5.4 Pro leads
Gemini 3.6 Flash: 57.3 (#40), GPT-5.4 Pro: 72.4 (#22)
| Benchmark | Gemini 3.6 Flash | GPT-5.4 Pro |
|---|---|---|
| FrontierMath (Tiers 1-3) | 58.9% | 82.5% |
| FrontierMath Tier 4 | 22% | 58.5% |
| MathArena Final-Answer Competitions | 70.8% | — |
| OTIS Mock AIME 2024-2025 | 94.2% | — |
| ProofBench | 36% | — |
| LMArena Math | 1505 | — |
| FrontierMath (Feb 2025 set) | — | 50% |
| FrontierMath Tier 4 (v1) | — | 37.5% |
Knowledge Too close to call
Gemini 3.6 Flash: 67.8 (#8), GPT-5.4 Pro: 68.3 (#7)
| Benchmark | Gemini 3.6 Flash | GPT-5.4 Pro |
|---|---|---|
| GPQA Diamond | 94.1% | 94.6% |
| SimpleQA Verified | 66.2% | 46.3% |
| Humanity's Last Exam | — | 44.3% |
| Vectara Hallucination Rate | — | 8.3% |
| LMArena Expert | 1488 | — |
Multimodal Not comparable
Gemini 3.6 Flash: 38.5 (#64), GPT-5.4 Pro: —
| Benchmark | Gemini 3.6 Flash | GPT-5.4 Pro |
|---|---|---|
| LMArena Vision | 1298 | — |
| Blueprint-Bench 2 | 31.2% | — |
| Furniture Assembly | 23.3% | — |
| LMArena Document | 1456 | — |
Multilingual Not comparable
Gemini 3.6 Flash: 56.5 (#19), GPT-5.4 Pro: —
| Benchmark | Gemini 3.6 Flash | GPT-5.4 Pro |
|---|---|---|
| LMArena Non-English | 1469 | — |
| LMArena Chinese | 1531 | — |
| LMArena French | 1504 | — |
| LMArena German | 1478 | — |
| LMArena Japanese | 1476 | — |
| LMArena Korean | 1431 | — |
| LMArena Russian | 1487 | — |
| LMArena Spanish | 1475 | — |
Instruction Following Not comparable
Gemini 3.6 Flash: 77.0 (#33), GPT-5.4 Pro: —
| Benchmark | Gemini 3.6 Flash | GPT-5.4 Pro |
|---|---|---|
| LMArena Instruction Following | 1466 | — |
Long Context Not comparable
Gemini 3.6 Flash: 45.1 (#50), GPT-5.4 Pro: —
| Benchmark | Gemini 3.6 Flash | GPT-5.4 Pro |
|---|---|---|
| LMArena Longer Query | 1474 | — |
Writing & Preference Not comparable
Gemini 3.6 Flash: 68.2 (#27), GPT-5.4 Pro: —
| Benchmark | Gemini 3.6 Flash | GPT-5.4 Pro |
|---|---|---|
| LMArena Text | 1479 | — |
| LMArena Creative Writing | 1465 | — |
| EQ-Bench Creative Writing | 1604 | — |
| LMArena Multi-Turn | 1481 | — |
Frequently asked questions
Is Gemini 3.6 Flash better than GPT-5.4 Pro?
GPT-5.4 Pro is the stronger model overall, scoring 58.9 to 54.1 on the Noometry Index. Gemini 3.6 Flash costs 45× less per token, which makes it the better buy when GPT-5.4 Pro's lead doesn't matter for your workload.
Which is cheaper, Gemini 3.6 Flash or GPT-5.4 Pro?
Gemini 3.6 Flash is cheaper. It lists at $0.75 per million input tokens and $3.75 per million output tokens; GPT-5.4 Pro lists at $30 and $180.
Is Gemini 3.6 Flash or GPT-5.4 Pro better for coding?
Gemini 3.6 Flash scores higher on coding benchmarks: 50.0 versus 43.2 in the Noometry coding category.
Which has the bigger context window?
GPT-5.4 Pro does, with 1.05M tokens against 1.05M.
How many benchmarks do Gemini 3.6 Flash and GPT-5.4 Pro share?
10 benchmarks have published results for both models. Gemini 3.6 Flash has 46 scored results on Noometry and GPT-5.4 Pro has 16.