Model comparison
DeepSeek V4.1 Flash vs DeepSeek V4 Pro
DeepSeek V4 Pro is the stronger model overall, scoring 54.3 to 52.8 on the Noometry Index. DeepSeek V4.1 Flash costs 3.8× less per token, which makes it the better buy when DeepSeek V4 Pro's lead doesn't matter for your workload.
Last verified . 34 shared benchmarks.
DeepSeekSummary
- They share 34 benchmarks with published results for both. DeepSeek V4.1 Flash scores higher in 5 categories and DeepSeek V4 Pro in 4 categories; 5 gaps are clear of the uncertainty.
- The widest gap is in reasoning, where DeepSeek V4 Pro leads 56.5 to 50.2.
- The biggest single-benchmark swing is APEX-Agents: 39.5% for DeepSeek V4.1 Flash and 47.3% for DeepSeek V4 Pro.
- DeepSeek V4.1 Flash is cheaper at $0.15 / $0.60 per million input/output tokens, against $0.66 / $1.98 for DeepSeek V4 Pro.
Side by side
| DeepSeek V4.1 Flash | DeepSeek V4 Pro | |
|---|---|---|
| Provider | DeepSeek | DeepSeek |
| Noometry Index | 52.8 | 54.3 |
| Released | 2026-09-09 | 2026-04-24 |
| Weights | Open | Open |
| Context window | 1M | 1M |
| Max output | 393K | 393K |
| Input $ / M tokens | $0.15 | $0.66 |
| Output $ / M tokens | $0.60 | $1.98 |
| Results tracked | 37 | 48 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding Too close to call
DeepSeek V4.1 Flash: 52.9 (#32), DeepSeek V4 Pro: 52.4 (#34)
| Benchmark | DeepSeek V4.1 Flash | DeepSeek V4 Pro |
|---|---|---|
| LMArena WebDev | 1619 | 1582 |
| SciCode | 51.9% | 51% |
| LMArena Coding | 1506 | 1470 |
| ALE-Bench | 1,092 | 1,403 |
| SWE-bench Verified | — | 77.6% |
| FrontierCode | — | 28.6% |
| WeirdML | — | 66.2% |
Agentic & Tool Use DeepSeek V4 Pro leads
DeepSeek V4.1 Flash: 31.2 (#69), DeepSeek V4 Pro: 32.8 (#58)
| Benchmark | DeepSeek V4.1 Flash | DeepSeek V4 Pro |
|---|---|---|
| APEX-Agents | 39.5% | 47.3% |
| GDP.pdf | 19.8% | — |
| Vending-Bench 2 | — | 3,285 |
Reasoning DeepSeek V4 Pro leads
DeepSeek V4.1 Flash: 50.2 (#36), DeepSeek V4 Pro: 56.5 (#24)
| Benchmark | DeepSeek V4.1 Flash | DeepSeek V4 Pro |
|---|---|---|
| NYT Connections (extended) | 89.6% | 91.3% |
| CritPt | 14.3% | 18% |
| LMArena Hard Prompts | 1483 | 1461 |
| Mystery Game Puzzles | 43% | 43% |
| DTBench | 89.9% | 93.9% |
| LMCA | 47% | 45.5% |
| Surface Evolver Bench | 46.3% | 40% |
| Epoch Capabilities Index | 154.9 | 155.31 |
| ARC-AGI-2 | — | 61.3% |
| Kagi LLM Benchmark | — | 53.5% |
| ARC-AGI-1 | — | 90.5% |
| Chess Puzzles | — | 47% |
| ForecastBench | — | 56.1 |
Math DeepSeek V4.1 Flash leads
DeepSeek V4.1 Flash: 66.7 (#25), DeepSeek V4 Pro: 64.8 (#30)
| Benchmark | DeepSeek V4.1 Flash | DeepSeek V4 Pro |
|---|---|---|
| FrontierMath (Tiers 1-3) | 67.4% | 64.6% |
| FrontierMath Tier 4 | 26.8% | 26.8% |
| OTIS Mock AIME 2024-2025 | 98.3% | 98.6% |
| ProofBench | 54% | 50% |
| LMArena Math | 1477 | 1455 |
| MathArena Final-Answer Competitions | — | 76.6% |
Knowledge DeepSeek V4 Pro leads
DeepSeek V4.1 Flash: 57.9 (#38), DeepSeek V4 Pro: 59.5 (#31)
| Benchmark | DeepSeek V4.1 Flash | DeepSeek V4 Pro |
|---|---|---|
| GPQA Diamond | 89.8% | 91.7% |
| LMArena Expert | 1506 | 1464 |
| SimpleQA Verified | — | 52.9% |
| Vectara Hallucination Rate | — | 8.6% |
Multimodal Not comparable
DeepSeek V4.1 Flash: 39.1 (#61), DeepSeek V4 Pro: —
| Benchmark | DeepSeek V4.1 Flash | DeepSeek V4 Pro |
|---|---|---|
| LMArena Vision | 1277 | — |
| Furniture Assembly | 34.2% | — |
Multilingual Too close to call
DeepSeek V4.1 Flash: 55.0 (#35), DeepSeek V4 Pro: 54.4 (#45)
| Benchmark | DeepSeek V4.1 Flash | DeepSeek V4 Pro |
|---|---|---|
| LMArena Non-English | 1448 | 1439 |
| LMArena Chinese | 1497 | 1486 |
| LMArena French | 1452 | 1472 |
| LMArena German | 1484 | 1458 |
| LMArena Japanese | 1412 | 1445 |
| LMArena Korean | 1452 | 1447 |
| LMArena Russian | 1471 | 1453 |
| LMArena Spanish | 1459 | 1458 |
Instruction Following DeepSeek V4.1 Flash leads
DeepSeek V4.1 Flash: 77.3 (#26), DeepSeek V4 Pro: 76.1 (#47)
| Benchmark | DeepSeek V4.1 Flash | DeepSeek V4 Pro |
|---|---|---|
| LMArena Instruction Following | 1474 | 1448 |
Long Context Too close to call
DeepSeek V4.1 Flash: 45.2 (#47), DeepSeek V4 Pro: 45.0 (#51)
| Benchmark | DeepSeek V4.1 Flash | DeepSeek V4 Pro |
|---|---|---|
| LMArena Longer Query | 1475 | 1458 |
| CL-bench Life | — | 13.5% |
Writing & Preference Too close to call
DeepSeek V4.1 Flash: 65.4 (#48), DeepSeek V4 Pro: 65.5 (#46)
| Benchmark | DeepSeek V4.1 Flash | DeepSeek V4 Pro |
|---|---|---|
| LMArena Text | 1462 | 1451 |
| LMArena Creative Writing | 1435 | 1446 |
| EQ-Bench Creative Writing | 1540 | 1553 |
| LMArena Multi-Turn | 1457 | 1467 |
| EQ-Bench 4 | — | 1166 |
Frequently asked questions
Is DeepSeek V4.1 Flash better than DeepSeek V4 Pro?
DeepSeek V4 Pro is the stronger model overall, scoring 54.3 to 52.8 on the Noometry Index. DeepSeek V4.1 Flash costs 3.8× less per token, which makes it the better buy when DeepSeek V4 Pro's lead doesn't matter for your workload.
Which is cheaper, DeepSeek V4.1 Flash or DeepSeek V4 Pro?
DeepSeek V4.1 Flash is cheaper. It lists at $0.15 per million input tokens and $0.60 per million output tokens; DeepSeek V4 Pro lists at $0.66 and $1.98.
Is DeepSeek V4.1 Flash or DeepSeek V4 Pro better for coding?
They score almost the same on coding (52.9 vs 52.4); test both on your own repository before choosing.
Which has the bigger context window?
Both accept 1M tokens.
How many benchmarks do DeepSeek V4.1 Flash and DeepSeek V4 Pro share?
34 benchmarks have published results for both models. DeepSeek V4.1 Flash has 37 scored results on Noometry and DeepSeek V4 Pro has 48.