DeepSeek V3.2 Exp vs Gemini 3.6 Flash
Gemini 3.6 Flash leads DeepSeek V3.2 Exp on the Artificial Analysis Intelligence Index, 34.0 to 16.6, and DeepSeek V3.2 Exp costs 9.1x less per output token ($0.41 vs $3.75 per million). Updated October 22, 2026; 44 benchmarks have a result for both.
DeepSeek V3.2 Exp or Gemini 3.6 Flash: which to pick
Only Gemini 3.6 Flash has a coding score, 62.5 on the Vibe Coding Index; DeepSeek V3.2 Exp has none, so the two cannot be ranked on coding.
Pick DeepSeek V3.2 Exp for output tokens 9.1x cheaper ($0.41 vs $3.75 per million).
- Shared results
- 44
- DeepSeek V3.2 Exp only
- 15
- Gemini 3.6 Flash only
- 65
- Like-for-like categories
- 7 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
No call
Only Gemini 3.6 Flash has a coding score, so no call is made.
One side not measured
Intelligence
AA Intelligence Index
Gemini 3.6 Flash
Gemini 3.6 Flash leads, 34.0 against 16.6.
Same board, both measured
Agentic work
AA Agentic Index
No call
Only Gemini 3.6 Flash has a AA Agentic Index result, 29.0.
Directional only
Reasoning
GPQA Diamond
No call
Only Gemini 3.6 Flash has a GPQA Diamond result, 93.4%.
Directional only
Knowledge
MMLU Pro
No call
Only Gemini 3.6 Flash has a MMLU Pro result, 89.3%.
Directional only
Web design
Design Arena: websites
Gemini 3.6 Flash
Gemini 3.6 Flash leads, 1300 against 1178.
Same board, both measured
Long documents
Listed context window
Gemini 3.6 Flash
Gemini 3.6 Flash takes 1.05M tokens against 164K.
Listed by OpenRouter
Everyday cost
1K input + 500 output tokens
DeepSeek V3.2 Exp
DeepSeek V3.2 Exp $0.0005, Gemini 3.6 Flash $0.0026 per turn.
List prices
Cost per solved task
One SWE-bench Verified task, solved
No call
Only Gemini 3.6 Flash has the number, $1.49.
Directional only
At a glance
- Coding index
- DeepSeek V3.2 Exp not benchmarked, Gemini 3.6 Flash 62.5 from 4 benchmarks.
- Widest category
- Coding, 23.4 points apart on average over 5 shared benchmarks.
- Widest single gap
- LMArena WebDev: HTML: DeepSeek V3.2 Exp 1281 against Gemini 3.6 Flash 1547, 44.4 points apart on a 0 to 100 scale.
- Output price
- DeepSeek V3.2 Exp $0.41 against Gemini 3.6 Flash $3.75 per million output tokens, 9.1x apart.
- Context window
- DeepSeek V3.2 Exp takes 164K tokens, Gemini 3.6 Flash 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 | DeepSeek V3.2 Exp | Bars | Gemini 3.6 Flash | Ties |
|---|---|---|---|---|
| Coding | 0 | 5 | ||
| Intelligence | 0 | 2 | ||
| Reasoning | 0 | 5 | ||
| Math | 0 | 2 | ||
| Knowledge | 0 | 4 | ||
| Web design | 0 | 6 | ||
| General | 0 | 20 | ||
| All shared boards | 0 | 44 |
CodingLike-for-likeSix of these boards feed the Vibe Coding Index05
| Benchmark | DeepSeek V3.2 Exp | Gemini 3.6 Flash | Leader |
|---|---|---|---|
| Aider polyglot | 74.2% | Not measured | Only DeepSeek V3.2 Exp measured |
| LMArena WebDev | 1272 | 1538 | Gemini 3.6 Flash |
| AA Coding Index | Not measured | 69.2 | Only Gemini 3.6 Flash measured |
| LiveCodeBench (Vals AI) | Not measured | 88.1% | Only Gemini 3.6 Flash measured |
| SWE-bench Verified (Vals AI) | Not measured | 79.6% | Only Gemini 3.6 Flash measured |
| Vibe Code Bench (Vals AI) | Not measured | 64.0% | Only Gemini 3.6 Flash measured |
| Code migration (Vals AI) | Not measured | 30.9% | Only Gemini 3.6 Flash measured |
| ProgramBench (Vals AI) | Not measured | 0.0% | Only Gemini 3.6 Flash measured |
| IOI (Vals AI) | Not measured | 35.1% | Only Gemini 3.6 Flash measured |
| Terminal-Bench 2.1 (Vals AI) | Not measured | 73.8% | Only Gemini 3.6 Flash measured |
| CyberBench (Vals AI) | Not measured | 45.4% | Only Gemini 3.6 Flash measured |
| LiveCodeBench (Artificial Analysis run) | 78.9% | Not measured | Only DeepSeek V3.2 Exp measured |
| SciCode | Not measured | 53.4% | Only Gemini 3.6 Flash measured |
| Terminal-Bench Hard (Artificial Analysis run) | 31.1% | Not measured | Only DeepSeek V3.2 Exp measured |
| FrontierCode | Not measured | 34.4% | Only Gemini 3.6 Flash measured |
| WeirdML | 39.5% | 56.1% | Gemini 3.6 Flash |
| ALE-Bench | Not measured | 715.5 | Only Gemini 3.6 Flash measured |
| LMArena Text: coding | 1438 | 1491 | Gemini 3.6 Flash |
| LMArena Text: software and IT | 1438 | 1492 | Gemini 3.6 Flash |
| LMArena WebDev: React | Not measured | 1528 | Only Gemini 3.6 Flash measured |
| LMArena WebDev: HTML | 1281 | 1547 | Gemini 3.6 Flash |
| LMArena WebDev: from an image | Not measured | 1545 | Only Gemini 3.6 Flash measured |
| SciCode | Not measured | 10.8% | Only Gemini 3.6 Flash measured |
| Neither model has a result on the other 16 boards here | |||
IntelligenceLike-for-likePublic boards, shown as published, kept out of the index02
| Benchmark | DeepSeek V3.2 Exp | Gemini 3.6 Flash | Leader |
|---|---|---|---|
| AA Intelligence Index | 16.6 | 34.0 | Gemini 3.6 Flash |
| Epoch Capabilities Index | 144.9 | 154.5 | Gemini 3.6 Flash |
Agentic workDirectional onlyPublic boards, shown as published, kept out of the index19 boards, none shared
| Benchmark | DeepSeek V3.2 Exp | Gemini 3.6 Flash | Leader |
|---|---|---|---|
| AA Agentic Index | Not measured | 29.0 | Only Gemini 3.6 Flash measured |
| Design Arena: agents building web apps | Not measured | 1197 | Only Gemini 3.6 Flash measured |
| Design Arena: agents building full-stack apps | Not measured | 1181 | Only Gemini 3.6 Flash measured |
| Design Arena: agents building mobile apps | Not measured | 1198 | Only Gemini 3.6 Flash measured |
| Design Arena: agents building Android apps | Not measured | 1204 | Only Gemini 3.6 Flash measured |
| Design Arena: agents building games | Not measured | 1180 | Only Gemini 3.6 Flash measured |
| Design Arena: HTML slides | Not measured | 1099 | Only Gemini 3.6 Flash measured |
| Design Arena: PowerPoint slides | Not measured | 1121 | Only Gemini 3.6 Flash measured |
| Design Arena: spreadsheets | Not measured | 1071 | Only Gemini 3.6 Flash measured |
| Finance Agent (Vals AI) | Not measured | 56.3% | Only Gemini 3.6 Flash measured |
| Tax Agent Bench (Vals AI) | Not measured | 18.0% | Only Gemini 3.6 Flash measured |
| Legal Research Bench (Vals AI) | Not measured | 25.0% | Only Gemini 3.6 Flash measured |
| Excel Modeling Benchmark (Vals AI) | Not measured | 65.4% | Only Gemini 3.6 Flash measured |
| BioMysteryBench (Vals AI) | Not measured | 58.5% | Only Gemini 3.6 Flash measured |
| Terminal-Bench-Science (Vals AI) | Not measured | 4.3% | Only Gemini 3.6 Flash measured |
| τ²-Bench Telecom | 33.9% | Not measured | Only DeepSeek V3.2 Exp measured |
| APEX-Agents | Not measured | 46.9% | Only Gemini 3.6 Flash measured |
| TheAgentCompany | 42.9% | Not measured | Only DeepSeek V3.2 Exp measured |
| Berkeley Function Calling Leaderboard (BFCL) | 56.7% | Not measured | Only DeepSeek V3.2 Exp measured |
| Neither model has a result on the other 22 boards here | |||
ReasoningLike-for-likePublic boards, shown as published, kept out of the index05
| Benchmark | DeepSeek V3.2 Exp | Gemini 3.6 Flash | Leader |
|---|---|---|---|
| GPQA Diamond | Not measured | 93.4% | Only Gemini 3.6 Flash measured |
| MMMU Pro (Vals AI) | Not measured | 88.4% | Only Gemini 3.6 Flash measured |
| GPQA Diamond (Artificial Analysis run) | 79.7% | 92.8% | Gemini 3.6 Flash |
| Humanity's Last Exam (Artificial Analysis run) | 14.9% | 40.8% | Gemini 3.6 Flash |
| GPQA Diamond (Epoch AI run) | Not measured | 94.1% | Only Gemini 3.6 Flash measured |
| ARC-AGI | Not measured | 91.2% | Only Gemini 3.6 Flash measured |
| ARC-AGI-2 | Not measured | 60.4% | Only Gemini 3.6 Flash measured |
| Chess puzzles | Not measured | 42.0% | Only Gemini 3.6 Flash measured |
| Mystery game puzzles | Not measured | 30.0% | Only Gemini 3.6 Flash measured |
| LMArena Text: hard prompts | 1429 | 1486 | Gemini 3.6 Flash |
| LMArena Text: expert | 1422 | 1490 | Gemini 3.6 Flash |
| LMArena Text: hard prompts in English | 1444 | 1484 | Gemini 3.6 Flash |
| LMArena Vision: diagrams | Not measured | 1314 | Only Gemini 3.6 Flash measured |
| LMArena Vision: homework | Not measured | 1314 | Only Gemini 3.6 Flash measured |
| Kagi LLM Benchmark | 52.2% | Not measured | Only DeepSeek V3.2 Exp measured |
| Neither model has a result on the other 7 boards here | |||
MathLike-for-likePublic boards, shown as published, kept out of the index02
| Benchmark | DeepSeek V3.2 Exp | Gemini 3.6 Flash | Leader |
|---|---|---|---|
| AA Math Index | 87.7 | Not measured | Only DeepSeek V3.2 Exp measured |
| AIME 2025 (Artificial Analysis run) | 87.7% | Not measured | Only DeepSeek V3.2 Exp measured |
| OTIS Mock AIME | Not measured | 94.2% | Only Gemini 3.6 Flash measured |
| FrontierMath, tiers 1 to 3 | Not measured | 58.9% | Only Gemini 3.6 Flash measured |
| FrontierMath, tier 4 | Not measured | 22.0% | Only Gemini 3.6 Flash measured |
| ProofBench | Not measured | 36.0% | Only Gemini 3.6 Flash measured |
| LMArena Text: math | 1422 | 1508 | Gemini 3.6 Flash |
| LMArena Text: mathematical work | 1433 | 1492 | Gemini 3.6 Flash |
| Neither model has a result on the other 5 boards here | |||
KnowledgeLike-for-likePublic boards, shown as published, kept out of the index04
| Benchmark | DeepSeek V3.2 Exp | Gemini 3.6 Flash | Leader |
|---|---|---|---|
| MMLU Pro | Not measured | 89.3% | Only Gemini 3.6 Flash measured |
| LegalBench (Vals AI) | Not measured | 86.7% | Only Gemini 3.6 Flash measured |
| CorpFin (Vals AI) | Not measured | 63.3% | Only Gemini 3.6 Flash measured |
| TaxEval (Vals AI) | Not measured | 74.9% | Only Gemini 3.6 Flash measured |
| MortgageTax (Vals AI) | Not measured | 67.6% | Only Gemini 3.6 Flash measured |
| MedCode (Vals AI) | Not measured | 53.1% | Only Gemini 3.6 Flash measured |
| MMLU Pro (Artificial Analysis run) | 85.0% | Not measured | Only DeepSeek V3.2 Exp measured |
| SimpleQA Verified | Not measured | 66.2% | Only Gemini 3.6 Flash measured |
| LMArena Text: science | 1443 | 1489 | Gemini 3.6 Flash |
| LMArena Text: medicine and healthcare | 1468 | 1469 | Gemini 3.6 Flash |
| LMArena Text: legal and government | 1419 | 1479 | Gemini 3.6 Flash |
| LMArena Text: business and finance | 1412 | 1464 | Gemini 3.6 Flash |
| FACTS Benchmark Suite | Not measured | 65.3% | Only Gemini 3.6 Flash measured |
| Neither model has a result on the other 5 boards here | |||
Web designLike-for-likePublic boards, shown as published, kept out of the index06
| Benchmark | DeepSeek V3.2 Exp | Gemini 3.6 Flash | Leader |
|---|---|---|---|
| Design Arena: websites | 1178 | 1300 | Gemini 3.6 Flash |
| Design Arena: UI components | 1164 | 1294 | Gemini 3.6 Flash |
| Design Arena: game dev | 1144 | 1263 | Gemini 3.6 Flash |
| Design Arena: data viz | 1152 | 1289 | Gemini 3.6 Flash |
| Design Arena: 3D | 1148 | 1271 | Gemini 3.6 Flash |
| Design Arena: all code categories | 1170 | 1289 | Gemini 3.6 Flash |
| Design Arena: SVG | 1031 | Not measured | Only DeepSeek V3.2 Exp measured |
| Design Arena: ASCII art | Not measured | 1281 | Only Gemini 3.6 Flash measured |
| Neither model has a result on the other 1 board here | |||
GeneralLike-for-likePublic boards, shown as published, kept out of the index020
| Benchmark | DeepSeek V3.2 Exp | Gemini 3.6 Flash | Leader |
|---|---|---|---|
| Public Benefits Bench (Vals AI) | Not measured | 56.6% | Only Gemini 3.6 Flash measured |
| SAGE (Vals AI) | Not measured | 48.7% | Only Gemini 3.6 Flash measured |
| MedScribe (Vals AI) | Not measured | 79.7% | Only Gemini 3.6 Flash measured |
| Vals Multimodal Index | Not measured | 65.1% | Only Gemini 3.6 Flash measured |
| IFBench | 54.1% | Not measured | Only DeepSeek V3.2 Exp measured |
| Long context reasoning (AA-LCR) | 72.3% | 80.0% | Gemini 3.6 Flash |
| DTBench | 87.7% | 95.5% | Gemini 3.6 Flash |
| LMCA | 29.1% | 44.9% | Gemini 3.6 Flash |
| Fiction.LiveBench | 83.3% | Not measured | Only DeepSeek V3.2 Exp measured |
| CL-bench | 13.2% | Not measured | Only DeepSeek V3.2 Exp measured |
| Furniture assembly | Not measured | 23.3% | Only Gemini 3.6 Flash measured |
| CL-bench Life | 9.5% | Not measured | Only DeepSeek V3.2 Exp measured |
| LMArena Text | 1425 | 1480 | Gemini 3.6 Flash |
| LMArena Text: instruction following | 1401 | 1467 | Gemini 3.6 Flash |
| LMArena Text: creative writing | 1404 | 1466 | Gemini 3.6 Flash |
| LMArena Vision | Not measured | 1299 | Only Gemini 3.6 Flash measured |
| LMArena Text: English | 1441 | 1482 | Gemini 3.6 Flash |
| LMArena Text: non-English | 1404 | 1470 | Gemini 3.6 Flash |
| LMArena Text: Chinese | 1461 | 1531 | Gemini 3.6 Flash |
| LMArena Text: German | 1440 | 1480 | Gemini 3.6 Flash |
| LMArena Text: Japanese | 1331 | 1475 | Gemini 3.6 Flash |
| LMArena Text: Korean | 1372 | 1429 | Gemini 3.6 Flash |
| LMArena Text: French | 1434 | 1503 | Gemini 3.6 Flash |
| LMArena Text: Spanish | 1441 | 1476 | Gemini 3.6 Flash |
| LMArena Text: Russian | 1412 | 1488 | Gemini 3.6 Flash |
| LMArena Text: Polish | 1435 | 1473 | Gemini 3.6 Flash |
| LMArena Text: multi-turn | 1422 | 1481 | Gemini 3.6 Flash |
| LMArena Text: longer queries | 1418 | 1475 | Gemini 3.6 Flash |
| LMArena Text: writing and language | 1407 | 1470 | Gemini 3.6 Flash |
| LMArena Text: entertainment and media | 1407 | 1456 | Gemini 3.6 Flash |
| LMArena Vision: English | Not measured | 1295 | Only Gemini 3.6 Flash measured |
| LMArena Vision: Chinese | Not measured | 1357 | Only Gemini 3.6 Flash measured |
| LMArena Vision: reading text | Not measured | 1312 | Only Gemini 3.6 Flash measured |
| LMArena Vision: humor | Not measured | 1289 | Only Gemini 3.6 Flash measured |
| LMArena Vision: creative writing | Not measured | 1278 | Only Gemini 3.6 Flash measured |
| LMArena Document | Not measured | 1456 | Only Gemini 3.6 Flash measured |
| Neither model has a result on the other 10 boards here | |||
Image generationNot comparablePublic boards, shown as published, kept out of the index0 boards, none shared
| Benchmark | DeepSeek V3.2 Exp | Gemini 3.6 Flash | 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 V3.2 Exp costs less
Repository review
50K input + 3K output tokens
DeepSeek V3.2 Exp costs less
Agent loop
200K input + 10K output tokens
DeepSeek V3.2 Exp 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 V3.2 Exp | Google Gemini 3.6 Flash |
|---|---|---|
| Maker | DeepSeek | |
| Weights | Open weights | No public weights listed |
| Released | September 29, 2025 | July 21, 2026 |
| Context window | 164K | 1.05M |
| Input $/Mtok | $0.27 | $0.75 |
| Output $/Mtok | $0.41 | $3.75 |
| Coding index | Not benchmarked | 62.5 |
| Independent measurements behind it | 1 | 4 |
| Cost per solved task | Not computable | $1.49 |
| Source | OpenRouter listing | OpenRouter listing |
Questions
Is DeepSeek V3.2 Exp better than Gemini 3.6 Flash for coding?
Only Gemini 3.6 Flash has a coding score, 62.5 on the Vibe Coding Index; DeepSeek V3.2 Exp has none, so the two cannot be ranked on coding. Pick DeepSeek V3.2 Exp for output tokens 9.1x cheaper ($0.41 vs $3.75 per million).
Which is cheaper, DeepSeek V3.2 Exp or Gemini 3.6 Flash?
DeepSeek V3.2 Exp costs $0.27 per million input tokens and $0.41 per million output tokens; Gemini 3.6 Flash costs $0.75 and $3.75. Cost per solved task is known for Gemini 3.6 Flash only.
Which has the bigger context window, DeepSeek V3.2 Exp or Gemini 3.6 Flash?
Gemini 3.6 Flash, with 1.05M tokens against 164K.
How many benchmarks is the DeepSeek V3.2 Exp vs Gemini 3.6 Flash comparison based on?
44 benchmarks have a result for both models as of October 22, 2026. DeepSeek V3.2 Exp's coding index rests on 1 independent benchmark and Gemini 3.6 Flash's on 4. Where only one model has a result, the page says so instead of naming a winner.