DeepSeek V4.1 Flash 1017 vs Gemini 3.5 Flash Lite
DeepSeek V4.1 Flash 1017 leads Gemini 3.5 Flash Lite on coding, 70.3 to 49.0 on the Vibe Coding Index, and DeepSeek V4.1 Flash 1017 costs 4.2x less per output token ($0.60 vs $2.50 per million). Updated October 22, 2026; 19 benchmarks have a result for both.
DeepSeek V4.1 Flash 1017 or Gemini 3.5 Flash Lite: which to pick
Pick DeepSeek V4.1 Flash 1017 for coding: it scores 70.3 on the Vibe Coding Index against 49.0 for Gemini 3.5 Flash Lite.
Gemini 3.5 Flash Lite is neither cheaper nor larger in context, and leads no benchmark both models were measured on.
- Shared results
- 19
- DeepSeek V4.1 Flash 1017 only
- 20
- Gemini 3.5 Flash Lite only
- 66
- Like-for-like categories
- 6 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
DeepSeek V4.1 Flash 1017
DeepSeek V4.1 Flash 1017 scores 70.3 against 49.0.
4 and 3 benchmarks behind the scores
Intelligence
AA Intelligence Index
DeepSeek V4.1 Flash 1017
DeepSeek V4.1 Flash 1017 leads, 39.5 against 22.2.
Same board, both measured
Agentic work
AA Agentic Index
No call
Only Gemini 3.5 Flash Lite has a AA Agentic Index result, 14.3.
Directional only
Reasoning
GPQA Diamond
No call
Only Gemini 3.5 Flash Lite has a GPQA Diamond result, 83.8%.
Directional only
Knowledge
MMLU Pro
No call
Only Gemini 3.5 Flash Lite has a MMLU Pro result, 85.8%.
Directional only
Web design
Design Arena: websites
No call
Only DeepSeek V4.1 Flash 1017 has a Design Arena: websites result, 1311.
Directional only
General
Vals Index
No call
Only DeepSeek V4.1 Flash 1017 has a Vals Index result, 61.3%.
Directional only
Long documents
Listed context window
No call
Both take 1.05M tokens.
Listed by OpenRouter
Everyday cost
1K input + 500 output tokens
DeepSeek V4.1 Flash 1017
DeepSeek V4.1 Flash 1017 $0.0004, Gemini 3.5 Flash Lite $0.0015 per turn.
List prices
Cost per solved task
One SWE-bench Verified task, solved
No call
Only Gemini 3.5 Flash Lite has the number, $0.41.
Directional only
At a glance
- Coding index
- DeepSeek V4.1 Flash 1017 70.3 from 4 benchmarks, Gemini 3.5 Flash Lite 49.0 from 3 benchmarks.
- Widest category
- Math, 40.0 points apart on average over 3 shared benchmarks.
- Widest single gap
- Vibe Code Bench (Vals AI): DeepSeek V4.1 Flash 1017 87.7% against Gemini 3.5 Flash Lite 37.2%, 50.6 points apart on a 0 to 100 scale.
- Output price
- DeepSeek V4.1 Flash 1017 $0.60 against Gemini 3.5 Flash Lite $2.50 per million output tokens, 4.2x apart.
- Context window
- Both take 1.05M tokens.
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 | DeepSeek V4.1 Flash 1017 | Bars | Gemini 3.5 Flash Lite | Ties |
|---|---|---|---|---|
| Coding | 6 | 0 | ||
| Intelligence | 1 | 0 | ||
| Agentic work | 4 | 0 | ||
| Reasoning | 4 | 0 | ||
| Math | 3 | 0 | ||
| General | 1 | 0 | ||
| All shared boards | 19 | 0 |
CodingLike-for-likeSix of these boards feed the Vibe Coding Index60
| Benchmark | DeepSeek V4.1 Flash 1017 | Gemini 3.5 Flash Lite | Leader |
|---|---|---|---|
| LiveBench coding | 96.0% | Not measured | Only DeepSeek V4.1 Flash 1017 measured |
| LMArena WebDev | 1664 | 1435 | DeepSeek V4.1 Flash 1017 |
| AA Coding Index | Not measured | 49.3 | Only Gemini 3.5 Flash Lite measured |
| LiveCodeBench (Vals AI) | Not measured | 79.0% | Only Gemini 3.5 Flash Lite measured |
| SWE-bench Verified (Vals AI) | Not measured | 75.0% | Only Gemini 3.5 Flash Lite measured |
| Terminal-Bench 4 (Vals AI) | 31.8% | Not measured | Only DeepSeek V4.1 Flash 1017 measured |
| Vibe Code Bench (Vals AI) | 87.7% | 37.2% | DeepSeek V4.1 Flash 1017 |
| Code migration (Vals AI) | 44.2% | 6.2% | DeepSeek V4.1 Flash 1017 |
| ProgramBench (Vals AI) | Not measured | 0.0% | Only Gemini 3.5 Flash Lite measured |
| Terminal-Bench 2.1 (Vals AI) | Not measured | 50.2% | Only Gemini 3.5 Flash Lite measured |
| CyberBench (Vals AI) | Not measured | 59.1% | Only Gemini 3.5 Flash Lite measured |
| SciCode | 55.9% | 41.3% | DeepSeek V4.1 Flash 1017 |
| WeirdML | Not measured | 39.0% | Only Gemini 3.5 Flash Lite measured |
| ALE-Bench | 1622.0 | 765.3 | DeepSeek V4.1 Flash 1017 |
| LMArena Text: coding | Not measured | 1453 | Only Gemini 3.5 Flash Lite measured |
| LMArena Text: software and IT | Not measured | 1451 | Only Gemini 3.5 Flash Lite measured |
| LMArena WebDev: React | 1667 | 1408 | DeepSeek V4.1 Flash 1017 |
| LMArena WebDev: HTML | 1687 | Not measured | Only DeepSeek V4.1 Flash 1017 measured |
| LMArena WebDev: from an image | Not measured | 1448 | Only Gemini 3.5 Flash Lite measured |
| SciCode | Not measured | 6.2% | Only Gemini 3.5 Flash Lite measured |
| Neither model has a result on the other 19 boards here | |||
IntelligenceLike-for-likePublic boards, shown as published, kept out of the index10
| Benchmark | DeepSeek V4.1 Flash 1017 | Gemini 3.5 Flash Lite | Leader |
|---|---|---|---|
| AA Intelligence Index | 39.5 | 22.2 | DeepSeek V4.1 Flash 1017 |
| Epoch Capabilities Index | Not measured | 144.9 | Only Gemini 3.5 Flash Lite measured |
Agentic workLike-for-likePublic boards, shown as published, kept out of the index40
| Benchmark | DeepSeek V4.1 Flash 1017 | Gemini 3.5 Flash Lite | Leader |
|---|---|---|---|
| AA Agentic Index | Not measured | 14.3 | Only Gemini 3.5 Flash Lite measured |
| Design Arena: spreadsheets | 1245 | Not measured | Only DeepSeek V4.1 Flash 1017 measured |
| LiveBench research | 79.5% | Not measured | Only DeepSeek V4.1 Flash 1017 measured |
| LiveBench browser | 76.0% | Not measured | Only DeepSeek V4.1 Flash 1017 measured |
| LiveBench long agentic | 50.9% | Not measured | Only DeepSeek V4.1 Flash 1017 measured |
| Finance Agent (Vals AI) | 57.0% | 47.4% | DeepSeek V4.1 Flash 1017 |
| Tax Agent Bench (Vals AI) | 44.4% | 7.2% | DeepSeek V4.1 Flash 1017 |
| Legal Research Bench (Vals AI) | 45.2% | 13.9% | DeepSeek V4.1 Flash 1017 |
| Harvey's Legal Agent Benchmark (Vals AI) | 14.2% | Not measured | Only DeepSeek V4.1 Flash 1017 measured |
| Excel Modeling Benchmark (Vals AI) | 61.8% | 43.2% | DeepSeek V4.1 Flash 1017 |
| APEX-Agents | Not measured | 29.3% | Only Gemini 3.5 Flash Lite measured |
| Neither model has a result on the other 30 boards here | |||
ReasoningLike-for-likePublic boards, shown as published, kept out of the index40
| Benchmark | DeepSeek V4.1 Flash 1017 | Gemini 3.5 Flash Lite | Leader |
|---|---|---|---|
| GPQA Diamond | Not measured | 83.8% | Only Gemini 3.5 Flash Lite measured |
| LiveBench reasoning | 89.1% | Not measured | Only DeepSeek V4.1 Flash 1017 measured |
| MMMU Pro (Vals AI) | Not measured | 83.6% | Only Gemini 3.5 Flash Lite measured |
| GPQA Diamond (Artificial Analysis run) | Not measured | 83.8% | Only Gemini 3.5 Flash Lite measured |
| Humanity's Last Exam (Artificial Analysis run) | 43.3% | 18.8% | DeepSeek V4.1 Flash 1017 |
| GPQA Diamond (Epoch AI run) | 93.2% | 83.3% | DeepSeek V4.1 Flash 1017 |
| ARC-AGI | Not measured | 53.5% | Only Gemini 3.5 Flash Lite measured |
| ARC-AGI-2 | Not measured | 10.3% | Only Gemini 3.5 Flash Lite measured |
| Chess puzzles | 46.0% | 21.0% | DeepSeek V4.1 Flash 1017 |
| Mystery game puzzles | 49.0% | 19.0% | DeepSeek V4.1 Flash 1017 |
| LMArena Text: hard prompts | Not measured | 1441 | Only Gemini 3.5 Flash Lite measured |
| LMArena Text: expert | Not measured | 1434 | Only Gemini 3.5 Flash Lite measured |
| LMArena Text: hard prompts in English | Not measured | 1440 | Only Gemini 3.5 Flash Lite measured |
| LMArena Vision: diagrams | Not measured | 1284 | Only Gemini 3.5 Flash Lite measured |
| LMArena Vision: homework | Not measured | 1291 | Only Gemini 3.5 Flash Lite measured |
| Neither model has a result on the other 7 boards here | |||
MathLike-for-likePublic boards, shown as published, kept out of the index30
| Benchmark | DeepSeek V4.1 Flash 1017 | Gemini 3.5 Flash Lite | Leader |
|---|---|---|---|
| OTIS Mock AIME | 99.4% | 71.1% | DeepSeek V4.1 Flash 1017 |
| FrontierMath, tiers 1 to 3 | 71.2% | 26.0% | DeepSeek V4.1 Flash 1017 |
| FrontierMath, tier 4 | 46.3% | 0.0% | DeepSeek V4.1 Flash 1017 |
| ProofBench | Not measured | 13.0% | Only Gemini 3.5 Flash Lite measured |
| LMArena Text: math | Not measured | 1434 | Only Gemini 3.5 Flash Lite measured |
| LMArena Text: mathematical work | Not measured | 1433 | Only Gemini 3.5 Flash Lite measured |
| Neither model has a result on the other 7 boards here | |||
KnowledgeDirectional onlyPublic boards, shown as published, kept out of the index12 boards, none shared
| Benchmark | DeepSeek V4.1 Flash 1017 | Gemini 3.5 Flash Lite | Leader |
|---|---|---|---|
| MMLU Pro | Not measured | 85.8% | Only Gemini 3.5 Flash Lite measured |
| LegalBench (Vals AI) | Not measured | 84.1% | Only Gemini 3.5 Flash Lite measured |
| CorpFin (Vals AI) | Not measured | 60.7% | Only Gemini 3.5 Flash Lite measured |
| TaxEval (Vals AI) | Not measured | 72.6% | Only Gemini 3.5 Flash Lite measured |
| MortgageTax (Vals AI) | Not measured | 68.7% | Only Gemini 3.5 Flash Lite measured |
| MedCode (Vals AI) | Not measured | 43.5% | Only Gemini 3.5 Flash Lite measured |
| SimpleQA Verified | 42.5% | Not measured | Only DeepSeek V4.1 Flash 1017 measured |
| LMArena Text: science | Not measured | 1451 | Only Gemini 3.5 Flash Lite measured |
| LMArena Text: medicine and healthcare | Not measured | 1438 | Only Gemini 3.5 Flash Lite measured |
| LMArena Text: legal and government | Not measured | 1431 | Only Gemini 3.5 Flash Lite measured |
| LMArena Text: business and finance | Not measured | 1423 | Only Gemini 3.5 Flash Lite measured |
| FACTS Benchmark Suite | Not measured | 53.5% | Only Gemini 3.5 Flash Lite measured |
| Neither model has a result on the other 6 boards here | |||
Web designDirectional onlyPublic boards, shown as published, kept out of the index8 boards, none shared
| Benchmark | DeepSeek V4.1 Flash 1017 | Gemini 3.5 Flash Lite | Leader |
|---|---|---|---|
| Design Arena: websites | 1311 | Not measured | Only DeepSeek V4.1 Flash 1017 measured |
| Design Arena: UI components | 1303 | Not measured | Only DeepSeek V4.1 Flash 1017 measured |
| Design Arena: game dev | 1357 | Not measured | Only DeepSeek V4.1 Flash 1017 measured |
| Design Arena: data viz | 1330 | Not measured | Only DeepSeek V4.1 Flash 1017 measured |
| Design Arena: 3D | 1358 | Not measured | Only DeepSeek V4.1 Flash 1017 measured |
| Design Arena: all code categories | 1332 | Not measured | Only DeepSeek V4.1 Flash 1017 measured |
| Design Arena: ASCII art | 1242 | Not measured | Only DeepSeek V4.1 Flash 1017 measured |
| LiveBench design | 83.4% | Not measured | Only DeepSeek V4.1 Flash 1017 measured |
| Neither model has a result on the other 1 board here | |||
GeneralLike-for-likePublic boards, shown as published, kept out of the index10
| Benchmark | DeepSeek V4.1 Flash 1017 | Gemini 3.5 Flash Lite | Leader |
|---|---|---|---|
| Vals Index | 61.3% | Not measured | Only DeepSeek V4.1 Flash 1017 measured |
| LiveBench data analysis | 79.5% | Not measured | Only DeepSeek V4.1 Flash 1017 measured |
| Public Benefits Bench (Vals AI) | Not measured | 51.8% | Only Gemini 3.5 Flash Lite measured |
| SAGE (Vals AI) | Not measured | 47.3% | Only Gemini 3.5 Flash Lite measured |
| MedScribe (Vals AI) | Not measured | 70.9% | Only Gemini 3.5 Flash Lite measured |
| Long context reasoning (AA-LCR) | 82.7% | 76.0% | DeepSeek V4.1 Flash 1017 |
| DTBench | Not measured | 83.5% | Only Gemini 3.5 Flash Lite measured |
| LMCA | Not measured | 37.6% | Only Gemini 3.5 Flash Lite measured |
| LMArena Text | Not measured | 1435 | Only Gemini 3.5 Flash Lite measured |
| LMArena Text: instruction following | Not measured | 1419 | Only Gemini 3.5 Flash Lite measured |
| LMArena Text: creative writing | Not measured | 1421 | Only Gemini 3.5 Flash Lite measured |
| LMArena Vision | Not measured | 1269 | Only Gemini 3.5 Flash Lite measured |
| LMArena Text: English | Not measured | 1438 | Only Gemini 3.5 Flash Lite measured |
| LMArena Text: non-English | Not measured | 1427 | Only Gemini 3.5 Flash Lite measured |
| LMArena Text: Chinese | Not measured | 1468 | Only Gemini 3.5 Flash Lite measured |
| LMArena Text: German | Not measured | 1454 | Only Gemini 3.5 Flash Lite measured |
| LMArena Text: Japanese | Not measured | 1426 | Only Gemini 3.5 Flash Lite measured |
| LMArena Text: Korean | Not measured | 1407 | Only Gemini 3.5 Flash Lite measured |
| LMArena Text: French | Not measured | 1451 | Only Gemini 3.5 Flash Lite measured |
| LMArena Text: Spanish | Not measured | 1441 | Only Gemini 3.5 Flash Lite measured |
| LMArena Text: Russian | Not measured | 1443 | Only Gemini 3.5 Flash Lite measured |
| LMArena Text: Polish | Not measured | 1437 | Only Gemini 3.5 Flash Lite measured |
| LMArena Text: multi-turn | Not measured | 1445 | Only Gemini 3.5 Flash Lite measured |
| LMArena Text: longer queries | Not measured | 1435 | Only Gemini 3.5 Flash Lite measured |
| LMArena Text: writing and language | Not measured | 1428 | Only Gemini 3.5 Flash Lite measured |
| LMArena Text: entertainment and media | Not measured | 1414 | Only Gemini 3.5 Flash Lite measured |
| LMArena Vision: English | Not measured | 1269 | Only Gemini 3.5 Flash Lite measured |
| LMArena Vision: Chinese | Not measured | 1311 | Only Gemini 3.5 Flash Lite measured |
| LMArena Vision: reading text | Not measured | 1280 | Only Gemini 3.5 Flash Lite measured |
| LMArena Vision: humor | Not measured | 1287 | Only Gemini 3.5 Flash Lite measured |
| LMArena Vision: creative writing | Not measured | 1260 | Only Gemini 3.5 Flash Lite measured |
| Neither model has a result on the other 15 boards here | |||
Image generationNot comparablePublic boards, shown as published, kept out of the index0 boards, none shared
| Benchmark | DeepSeek V4.1 Flash 1017 | Gemini 3.5 Flash Lite | 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
DeepSeek V4.1 Flash 1017 costs less
Repository review
50K input + 3K output tokens
DeepSeek V4.1 Flash 1017 costs less
Agent loop
200K input + 10K output tokens
DeepSeek V4.1 Flash 1017 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 | DeepSeek V4.1 Flash 1017 | Google Gemini 3.5 Flash Lite |
|---|---|---|
| Maker | DeepSeek | |
| Weights | Open weights | No public weights listed |
| Released | October 17, 2026 | July 21, 2026 |
| Context window | 1.05M | 1.05M |
| Input $/Mtok | $0.15 | $0.30 |
| Output $/Mtok | $0.60 | $2.50 |
| Coding index | 70.3 | 49.0 |
| Independent measurements behind it | 4 | 3 |
| Cost per solved task | Not computable | $0.41 |
| Source | OpenRouter listing | OpenRouter listing |
Questions
Is DeepSeek V4.1 Flash 1017 better than Gemini 3.5 Flash Lite for coding?
Pick DeepSeek V4.1 Flash 1017 for coding: it scores 70.3 on the Vibe Coding Index against 49.0 for Gemini 3.5 Flash Lite. Gemini 3.5 Flash Lite is neither cheaper nor larger in context, and leads no benchmark both models were measured on.
Which is cheaper, DeepSeek V4.1 Flash 1017 or Gemini 3.5 Flash Lite?
DeepSeek V4.1 Flash 1017 costs $0.15 per million input tokens and $0.60 per million output tokens; Gemini 3.5 Flash Lite costs $0.30 and $2.50. Cost per solved task is known for Gemini 3.5 Flash Lite only.
Which has the bigger context window, DeepSeek V4.1 Flash 1017 or Gemini 3.5 Flash Lite?
Neither: both take 1.05M tokens.
How many benchmarks is the DeepSeek V4.1 Flash 1017 vs Gemini 3.5 Flash Lite comparison based on?
19 benchmarks have a result for both models as of October 22, 2026. DeepSeek V4.1 Flash 1017's coding index rests on 4 independent benchmarks and Gemini 3.5 Flash Lite's on 3. Where only one model has a result, the page says so instead of naming a winner.