Gemini 3 Flash Preview vs GPT-5.4
GPT-5.4 leads Gemini 3 Flash Preview on coding, 64.4 to 43.7 on the Vibe Coding Index, and Gemini 3 Flash Preview costs 5.0x less per output token ($3.00 vs $15.00 per million). Updated October 22, 2026; 52 benchmarks have a result for both.
Gemini 3 Flash Preview or GPT-5.4: which to pick
Pick GPT-5.4 for coding: it scores 64.4 on the Vibe Coding Index against 43.7 for Gemini 3 Flash Preview.
Pick Gemini 3 Flash Preview for output tokens 5.0x cheaper ($3.00 vs $15.00 per million) and a lower cost per solved task ($0.37 against $1.02).
- Shared results
- 52
- Gemini 3 Flash Preview only
- 19
- GPT-5.4 only
- 58
- Like-for-like categories
- 8 of 9
Only benchmarks that tested both models count toward a pick; a result only one model has is shown but not compared.
Which one for your work
Each card is decided by one number both models carry. Where only one has it, the card says so and makes no call.
Coding
Vibe Coding Index, six public coding benchmarks
GPT-5.4
GPT-5.4 scores 64.4 against 43.7.
2 and 3 benchmarks behind the scores
Intelligence
AA Intelligence Index
GPT-5.4
GPT-5.4 leads, 39.0 against 26.3.
Same board, both measured
Reasoning
GPQA Diamond
GPT-5.4
GPT-5.4 leads, 91.7% against 87.9%.
Same board, both measured
Math
AIME
GPT-5.4
GPT-5.4 leads, 96.7% against 95.6%.
Same board, both measured
Knowledge
MMLU Pro
Gemini 3 Flash Preview
Gemini 3 Flash Preview leads, 88.6% against 87.5%.
Same board, both measured
Web design
Design Arena: websites
GPT-5.4
GPT-5.4 leads, 1198 against 1196.
Same board, both measured
Long documents
Listed context window
GPT-5.4
GPT-5.4 takes 1.05M tokens against 1.05M.
Listed by OpenRouter
Everyday cost
1K input + 500 output tokens
Gemini 3 Flash Preview
Gemini 3 Flash Preview $0.0020, GPT-5.4 $0.010 per turn.
List prices
Cost per solved task
One SWE-bench Verified task, solved
Gemini 3 Flash Preview
Gemini 3 Flash Preview $0.37, GPT-5.4 $1.02.
Published attempt cost and pass rate
At a glance
- Coding index
- Gemini 3 Flash Preview 43.7 from 2 benchmarks, GPT-5.4 64.4 from 3 benchmarks.
- Widest category
- Math, 20.7 points apart on average over 5 shared benchmarks.
- Widest single gap
- Vibe Code Bench (Vals AI): Gemini 3 Flash Preview 20.2% against GPT-5.4 67.4%, 47.2 points apart on a 0 to 100 scale.
- Output price
- Gemini 3 Flash Preview $3.00 against GPT-5.4 $15.00 per million output tokens, 5.0x apart.
- Context window
- Gemini 3 Flash Preview takes 1.05M tokens, GPT-5.4 1.05M.
Benchmark by benchmark
First the score: boards each model leads, per category, where both have a result. Then every board, a category to a fold; bars are each result on a 0 to 100 scale, so compare the two in a row, not one row with another.
| Category | Gemini 3 Flash Preview | Bars | GPT-5.4 | Ties |
|---|---|---|---|---|
| Coding | 1 | 9 | ||
| Intelligence | 0 | 1 | ||
| Agentic work | 6 | 2 | ||
| Reasoning | 1 | 8 | ||
| Math | 0 | 5 | ||
| Knowledge | 6 | 3 | ||
| Web design | 2 | 2 | ||
| General | 2 | 4 | ||
| All shared boards | 18 | 34 |
CodingLike-for-likeSix of these boards feed the Vibe Coding Index19
| Benchmark | Gemini 3 Flash Preview | GPT-5.4 | Leader |
|---|---|---|---|
| SWE-bench Verified | 75.8% | Not measured | Only Gemini 3 Flash Preview measured |
| LMArena WebDev | Not measured | 1464 | Only GPT-5.4 measured |
| AA Coding Index | Not measured | 71.1 | Only GPT-5.4 measured |
| LiveCodeBench (Vals AI) | 85.6% | 84.1% | Gemini 3 Flash Preview |
| SWE-bench Verified (Vals AI) | 75.0% | 78.2% | GPT-5.4 |
| Vibe Code Bench (Vals AI) | 20.2% | 67.4% | GPT-5.4 |
| Code migration (Vals AI) | 6.4% | 35.0% | GPT-5.4 |
| ProgramBench (Vals AI) | 0.0% | 0.5% | GPT-5.4 |
| Terminal-Bench 2.1 (Vals AI) | 53.9% | Not measured | Only Gemini 3 Flash Preview measured |
| SkillsBench (Vals AI) | Not measured | 51.7% | Only GPT-5.4 measured |
| LiveCodeBench (Artificial Analysis run) | 90.8% | Not measured | Only Gemini 3 Flash Preview measured |
| Terminal-Bench Hard (Artificial Analysis run) | 38.6% | 57.6% | GPT-5.4 |
| Terminal-Bench | 64.3% | 81.8% | GPT-5.4 |
| GSO-Bench | 9.8% | 31.4% | GPT-5.4 |
| WeirdML | 61.6% | 77.7% | GPT-5.4 |
| MirrorCode | Not measured | 15.6% | Only GPT-5.4 measured |
| ALE-Bench | 1367.2 | 1607.0 | GPT-5.4 |
| LMArena Text: coding | Not measured | 1496 | Only GPT-5.4 measured |
| LMArena Text: software and IT | Not measured | 1491 | Only GPT-5.4 measured |
| LMArena WebDev: React | Not measured | 1454 | Only GPT-5.4 measured |
| LMArena WebDev: HTML | Not measured | 1478 | Only GPT-5.4 measured |
| LMArena WebDev: from an image | Not measured | 1436 | Only GPT-5.4 measured |
| SciCode | 12.7% | Not measured | Only Gemini 3 Flash Preview measured |
| Neither model has a result on the other 16 boards here | |||
IntelligenceLike-for-likePublic boards, shown as published, kept out of the index01
| Benchmark | Gemini 3 Flash Preview | GPT-5.4 | Leader |
|---|---|---|---|
| AA Intelligence Index | 26.3 | 39.0 | GPT-5.4 |
| Epoch Capabilities Index | Not measured | 157.3 | Only GPT-5.4 measured |
Agentic workLike-for-likePublic boards, shown as published, kept out of the index62
| Benchmark | Gemini 3 Flash Preview | GPT-5.4 | Leader |
|---|---|---|---|
| Design Arena: agents building web apps | 1111 | 1059 | Gemini 3 Flash Preview |
| Design Arena: agents building full-stack apps | 1063 | 1017 | Gemini 3 Flash Preview |
| Design Arena: agents building mobile apps | 1090 | 1087 | Gemini 3 Flash Preview |
| Design Arena: agents building Android apps | 1051 | 1031 | Gemini 3 Flash Preview |
| Design Arena: agents building Godot games | 1161 | 1135 | Gemini 3 Flash Preview |
| Design Arena: PowerPoint slides | 1036 | Not measured | Only Gemini 3 Flash Preview measured |
| Finance Agent (Vals AI) | 42.5% | Not measured | Only Gemini 3 Flash Preview measured |
| Tax Agent Bench (Vals AI) | 11.8% | Not measured | Only Gemini 3 Flash Preview measured |
| Legal Research Bench (Vals AI) | 18.3% | Not measured | Only Gemini 3 Flash Preview measured |
| Excel Modeling Benchmark (Vals AI) | 26.0% | Not measured | Only Gemini 3 Flash Preview measured |
| τ²-Bench Telecom | 80.4% | 87.1% | GPT-5.4 |
| APEX-Agents | Not measured | 52.4% | Only GPT-5.4 measured |
| METR time horizons | Not measured | 74.3% | Only GPT-5.4 measured |
| BALROG | 48.1% | Not measured | Only Gemini 3 Flash Preview measured |
| DeepResearch Bench | 49.8% | 35.1% | Gemini 3 Flash Preview |
| PostTrainBench | Not measured | 19.0% | Only GPT-5.4 measured |
| Vending-Bench 2 | $3,635 | $6,144 | GPT-5.4 |
| Neither model has a result on the other 24 boards here | |||
ReasoningLike-for-likePublic boards, shown as published, kept out of the index18
| Benchmark | Gemini 3 Flash Preview | GPT-5.4 | Leader |
|---|---|---|---|
| GPQA Diamond | 87.9% | 91.7% | GPT-5.4 |
| MMMU Pro (Vals AI) | 87.6% | 87.5% | Gemini 3 Flash Preview |
| GPQA Diamond (Artificial Analysis run) | 89.8% | 92.0% | GPT-5.4 |
| Humanity's Last Exam (Artificial Analysis run) | 36.6% | 43.7% | GPT-5.4 |
| GPQA Diamond (Epoch AI run) | 89.4% | 93.3% | GPT-5.4 |
| ARC-AGI | 84.7% | 93.7% | GPT-5.4 |
| ARC-AGI-2 | 33.6% | 74.0% | GPT-5.4 |
| SimpleBench | 61.1% | Not measured | Only Gemini 3 Flash Preview measured |
| Chess puzzles | 39.0% | 48.0% | GPT-5.4 |
| Mystery game puzzles | 26.0% | 37.0% | GPT-5.4 |
| EBR-bench | Not measured | 25.4% | Only GPT-5.4 measured |
| VPCT | 72.6% | Not measured | Only Gemini 3 Flash Preview measured |
| LMArena Text: hard prompts | Not measured | 1485 | Only GPT-5.4 measured |
| LMArena Text: expert | Not measured | 1507 | Only GPT-5.4 measured |
| LMArena Text: hard prompts in English | Not measured | 1479 | Only GPT-5.4 measured |
| LMArena Vision: diagrams | Not measured | 1327 | Only GPT-5.4 measured |
| LMArena Vision: homework | Not measured | 1336 | Only GPT-5.4 measured |
| Kagi LLM Benchmark | Not measured | 63.8% | Only GPT-5.4 measured |
| Neither model has a result on the other 4 boards here | |||
MathLike-for-likePublic boards, shown as published, kept out of the index05
| Benchmark | Gemini 3 Flash Preview | GPT-5.4 | Leader |
|---|---|---|---|
| AIME | 95.6% | 96.7% | GPT-5.4 |
| MGSM (Vals AI) | 93.3% | Not measured | Only Gemini 3 Flash Preview measured |
| AA Math Index | 97.0 | Not measured | Only Gemini 3 Flash Preview measured |
| AIME 2025 (Artificial Analysis run) | 97.0% | Not measured | Only Gemini 3 Flash Preview measured |
| OTIS Mock AIME | 95.6% | 97.8% | GPT-5.4 |
| FrontierMath, tiers 1 to 3 | 51.2% | 78.6% | GPT-5.4 |
| FrontierMath, tier 4 | 17.1% | 49.0% | GPT-5.4 |
| ProofBench | 15.0% | 56.0% | GPT-5.4 |
| LMArena Text: math | Not measured | 1488 | Only GPT-5.4 measured |
| LMArena Text: mathematical work | Not measured | 1497 | Only GPT-5.4 measured |
| Neither model has a result on the other 3 boards here | |||
KnowledgeLike-for-likePublic boards, shown as published, kept out of the index63
| Benchmark | Gemini 3 Flash Preview | GPT-5.4 | Leader |
|---|---|---|---|
| MMLU Pro | 88.6% | 87.5% | Gemini 3 Flash Preview |
| LegalBench (Vals AI) | 86.9% | 86.0% | Gemini 3 Flash Preview |
| CaseLaw (Vals AI) | 55.8% | 63.8% | GPT-5.4 |
| CorpFin (Vals AI) | 66.4% | 65.3% | Gemini 3 Flash Preview |
| TaxEval (Vals AI) | 73.9% | 74.0% | GPT-5.4 |
| MortgageTax (Vals AI) | 68.7% | 68.3% | Gemini 3 Flash Preview |
| MedQA (Vals AI) | 95.8% | 96.1% | GPT-5.4 |
| MedCode (Vals AI) | 55.9% | 41.3% | Gemini 3 Flash Preview |
| MMLU Pro (Artificial Analysis run) | 89.0% | Not measured | Only Gemini 3 Flash Preview measured |
| SimpleQA Verified | 66.8% | 45.1% | Gemini 3 Flash Preview |
| GeoBench | 88.0% | Not measured | Only Gemini 3 Flash Preview measured |
| LMArena Text: science | Not measured | 1479 | Only GPT-5.4 measured |
| LMArena Text: medicine and healthcare | Not measured | 1458 | Only GPT-5.4 measured |
| LMArena Text: legal and government | Not measured | 1486 | Only GPT-5.4 measured |
| LMArena Text: business and finance | Not measured | 1474 | Only GPT-5.4 measured |
| FACTS Benchmark Suite | 63.4% | Not measured | Only Gemini 3 Flash Preview measured |
| Neither model has a result on the other 2 boards here | |||
Web designLike-for-likePublic boards, shown as published, kept out of the index22
| Benchmark | Gemini 3 Flash Preview | GPT-5.4 | Leader |
|---|---|---|---|
| Design Arena: websites | 1196 | 1198 | GPT-5.4 |
| Design Arena: UI components | Not measured | 1212 | Only GPT-5.4 measured |
| Design Arena: game dev | 1166 | 1194 | GPT-5.4 |
| Design Arena: data viz | Not measured | 1196 | Only GPT-5.4 measured |
| Design Arena: 3D | 1183 | 1065 | Gemini 3 Flash Preview |
| Design Arena: all code categories | 1189 | 1179 | Gemini 3 Flash Preview |
| Design Arena: SVG | Not measured | 1141 | Only GPT-5.4 measured |
| Design Arena: ASCII art | Not measured | 1188 | Only GPT-5.4 measured |
| Neither model has a result on the other 1 board here | |||
GeneralLike-for-likePublic boards, shown as published, kept out of the index24
| Benchmark | Gemini 3 Flash Preview | GPT-5.4 | Leader |
|---|---|---|---|
| SAGE (Vals AI) | 51.8% | 43.3% | Gemini 3 Flash Preview |
| MedScribe (Vals AI) | 69.9% | 77.5% | GPT-5.4 |
| Vals Multimodal Index | 52.2% | Not measured | Only Gemini 3 Flash Preview measured |
| IFBench | 78.0% | 73.9% | Gemini 3 Flash Preview |
| Long context reasoning (AA-LCR) | 78.0% | 82.0% | GPT-5.4 |
| DTBench | 89.1% | 94.4% | GPT-5.4 |
| LMCA | 43.1% | 52.0% | GPT-5.4 |
| CL-bench | Not measured | 27.9% | Only GPT-5.4 measured |
| Furniture assembly | Not measured | 37.5% | Only GPT-5.4 measured |
| CL-bench Life | Not measured | 21.7% | Only GPT-5.4 measured |
| LMArena Text | Not measured | 1469 | Only GPT-5.4 measured |
| LMArena Text: instruction following | Not measured | 1469 | Only GPT-5.4 measured |
| LMArena Text: creative writing | Not measured | 1439 | Only GPT-5.4 measured |
| LMArena Vision | Not measured | 1303 | Only GPT-5.4 measured |
| LMArena Text: English | Not measured | 1467 | Only GPT-5.4 measured |
| LMArena Text: non-English | Not measured | 1465 | Only GPT-5.4 measured |
| LMArena Text: Chinese | Not measured | 1520 | Only GPT-5.4 measured |
| LMArena Text: German | Not measured | 1472 | Only GPT-5.4 measured |
| LMArena Text: Japanese | Not measured | 1485 | Only GPT-5.4 measured |
| LMArena Text: Korean | Not measured | 1449 | Only GPT-5.4 measured |
| LMArena Text: French | Not measured | 1493 | Only GPT-5.4 measured |
| LMArena Text: Spanish | Not measured | 1452 | Only GPT-5.4 measured |
| LMArena Text: Russian | Not measured | 1480 | Only GPT-5.4 measured |
| LMArena Text: Polish | Not measured | 1472 | Only GPT-5.4 measured |
| LMArena Text: multi-turn | Not measured | 1482 | Only GPT-5.4 measured |
| LMArena Text: longer queries | Not measured | 1473 | Only GPT-5.4 measured |
| LMArena Text: writing and language | Not measured | 1458 | Only GPT-5.4 measured |
| LMArena Text: entertainment and media | Not measured | 1438 | Only GPT-5.4 measured |
| LMArena Vision: English | Not measured | 1301 | Only GPT-5.4 measured |
| LMArena Vision: Chinese | Not measured | 1354 | Only GPT-5.4 measured |
| LMArena Vision: reading text | Not measured | 1312 | Only GPT-5.4 measured |
| LMArena Vision: humor | Not measured | 1284 | Only GPT-5.4 measured |
| LMArena Vision: recognizing things | Not measured | 1251 | Only GPT-5.4 measured |
| LMArena Vision: creative writing | Not measured | 1283 | Only GPT-5.4 measured |
| LMArena Document | Not measured | 1471 | Only GPT-5.4 measured |
| Neither model has a result on the other 11 boards here | |||
Image generationNot comparablePublic boards, shown as published, kept out of the index0 boards, none shared
| Benchmark | Gemini 3 Flash Preview | GPT-5.4 | Leader |
|---|---|---|---|
| Neither model has a result on 5 boards here | |||
What each workload costs
Three fixed token mixes at each model's listed price, so per-token rates become a bill.
Chat turn
1K input + 500 output tokens
Gemini 3 Flash Preview costs less
Repository review
50K input + 3K output tokens
Gemini 3 Flash Preview costs less
Agent loop
200K input + 10K output tokens
Gemini 3 Flash Preview costs less
Listed OpenRouter rates, without cache or batch discounts. A workload that does not fit a model's context window would need splitting.
Specification differences
The leading value in a row is marked; a row with a missing value names no leader. Our two rows are marked in red.
| Specification | Google Gemini 3 Flash Preview | OpenAI GPT-5.4 |
|---|---|---|
| Maker | OpenAI | |
| Weights | No public weights listed | No public weights listed |
| Released | December 17, 2025 | March 5, 2026 |
| Context window | 1.05M | 1.05M |
| Input $/Mtok | $0.50 | $2.50 |
| Output $/Mtok | $3.00 | $15.00 |
| Coding index | 43.7 | 64.4 |
| Independent measurements behind it | 2 | 3 |
| Cost per solved task | $0.37 | $1.02 |
| Source | OpenRouter listing | OpenRouter listing |
Questions
Is Gemini 3 Flash Preview better than GPT-5.4 for coding?
Pick GPT-5.4 for coding: it scores 64.4 on the Vibe Coding Index against 43.7 for Gemini 3 Flash Preview. Pick Gemini 3 Flash Preview for output tokens 5.0x cheaper ($3.00 vs $15.00 per million) and a lower cost per solved task ($0.37 against $1.02).
Which is cheaper, Gemini 3 Flash Preview or GPT-5.4?
Gemini 3 Flash Preview costs $0.50 per million input tokens and $3.00 per million output tokens; GPT-5.4 costs $2.50 and $15.00. Per solved SWE-bench Verified task, Gemini 3 Flash Preview costs $0.37 and GPT-5.4 $1.02.
Which has the bigger context window, Gemini 3 Flash Preview or GPT-5.4?
GPT-5.4, with 1.05M tokens against 1.05M.
How many benchmarks is the Gemini 3 Flash Preview vs GPT-5.4 comparison based on?
52 benchmarks have a result for both models as of October 22, 2026. Gemini 3 Flash Preview's coding index rests on 2 independent benchmarks and GPT-5.4's on 3. Where only one model has a result, the page says so instead of naming a winner.