Model comparison
DeepSeek V4 Pro vs Mercury 2
DeepSeek V4 Pro is the stronger model overall, scoring 54.2 to 38.8 on the Noometry Index.
Last verified . 17 shared benchmarks.
DeepSeekSummary
- They share 17 benchmarks with published results for both. DeepSeek V4 Pro scores higher in 7 categories and Mercury 2 in 0 categories; 7 gaps are clear of the uncertainty.
- The widest gap is in reasoning, where DeepSeek V4 Pro leads 56.4 to 23.0.
- The biggest single-benchmark swing is WeirdML: 66.2% for DeepSeek V4 Pro and 43.2% for Mercury 2.
- Mercury 2 is cheaper at $0.25 / $0.75 per million input/output tokens, against $0.43 / $0.87 for DeepSeek V4 Pro.
- DeepSeek V4 Pro accepts more context: 1M tokens versus 128K.
- DeepSeek V4 Pro has downloadable open weights; the other is API-only.
Side by side
| DeepSeek V4 Pro | Mercury 2 | |
|---|---|---|
| Provider | DeepSeek | Inception |
| Noometry Index | 54.2 | 38.8 |
| Released | 2026-04-24 | 2026-02-20 |
| Weights | Open | Proprietary |
| Context window | 1M | 128K |
| Max output | 393K | 50K |
| Input $ / M tokens | $0.43 | $0.25 |
| Output $ / M tokens | $0.87 | $0.75 |
| Results tracked | 48 | 17 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding DeepSeek V4 Pro leads
DeepSeek V4 Pro: 52.2 (#31), Mercury 2: 33.2 (#250)
| Benchmark | DeepSeek V4 Pro | Mercury 2 |
|---|---|---|
| LMArena WebDev | 1582 | 1170 |
| SciCode | 51% | 38.7% |
| WeirdML | 66.2% | 43.2% |
| LMArena Coding | 1470 | 1391 |
| ALE-Bench | 1,403 | 785.58 |
| SWE-bench Verified | 77.6% | — |
| FrontierCode | 28.6% | — |
Agentic & Tool Use Not comparable
DeepSeek V4 Pro: 32.7 (#54), Mercury 2: —
| Benchmark | DeepSeek V4 Pro | Mercury 2 |
|---|---|---|
| APEX-Agents | 47.3% | — |
| Vending-Bench 2 | 3,285 | — |
Reasoning DeepSeek V4 Pro leads
DeepSeek V4 Pro: 56.4 (#22), Mercury 2: 23.0 (#170)
| Benchmark | DeepSeek V4 Pro | Mercury 2 |
|---|---|---|
| CritPt | 18% | 0.8% |
| LMArena Hard Prompts | 1461 | 1362 |
| ARC-AGI-2 | 61.3% | — |
| Kagi LLM Benchmark | 53.5% | — |
| NYT Connections (extended) | 91.3% | — |
| ARC-AGI-1 | 90.5% | — |
| Chess Puzzles | 47% | — |
| Mystery Game Puzzles | 43% | — |
| DTBench | 93.9% | — |
| LMCA | 45.5% | — |
| Surface Evolver Bench | 40% | — |
| Epoch Capabilities Index | 155.39 | — |
| ForecastBench | 56.1 | — |
Math Not comparable
DeepSeek V4 Pro: 64.6 (#24), Mercury 2: —
| Benchmark | DeepSeek V4 Pro | Mercury 2 |
|---|---|---|
| FrontierMath (Tiers 1-3) | 64.6% | — |
| FrontierMath Tier 4 | 26.8% | — |
| MathArena Final-Answer Competitions | 76.6% | — |
| OTIS Mock AIME 2024-2025 | 98.6% | — |
| ProofBench | 50% | — |
| LMArena Math | 1455 | — |
Knowledge DeepSeek V4 Pro leads
DeepSeek V4 Pro: 59.5 (#28), Mercury 2: 35.6 (#170)
| Benchmark | DeepSeek V4 Pro | Mercury 2 |
|---|---|---|
| Vectara Hallucination Rate | 8.6% | 12.3% |
| LMArena Expert | 1463 | 1358 |
| GPQA Diamond | 91.7% | — |
| SimpleQA Verified | 52.9% | — |
Multilingual DeepSeek V4 Pro leads
DeepSeek V4 Pro: 54.5 (#44), Mercury 2: 46.8 (#153)
| Benchmark | DeepSeek V4 Pro | Mercury 2 |
|---|---|---|
| LMArena Non-English | 1438 | 1331 |
| LMArena Chinese | 1486 | 1416 |
| LMArena Russian | 1452 | 1303 |
| LMArena French | 1473 | — |
| LMArena German | 1458 | — |
| LMArena Japanese | 1448 | — |
| LMArena Korean | 1452 | — |
| LMArena Spanish | 1457 | — |
Instruction Following DeepSeek V4 Pro leads
DeepSeek V4 Pro: 76.3 (#44), Mercury 2: 70.4 (#161)
| Benchmark | DeepSeek V4 Pro | Mercury 2 |
|---|---|---|
| LMArena Instruction Following | 1448 | 1328 |
Long Context DeepSeek V4 Pro leads
DeepSeek V4 Pro: 45.0 (#49), Mercury 2: 40.4 (#149)
| Benchmark | DeepSeek V4 Pro | Mercury 2 |
|---|---|---|
| LMArena Longer Query | 1459 | 1329 |
| CL-bench Life | 13.5% | — |
Writing & Preference DeepSeek V4 Pro leads
DeepSeek V4 Pro: 65.6 (#45), Mercury 2: 54.0 (#151)
| Benchmark | DeepSeek V4 Pro | Mercury 2 |
|---|---|---|
| LMArena Text | 1451 | 1355 |
| LMArena Creative Writing | 1446 | 1289 |
| LMArena Multi-Turn | 1467 | 1358 |
| EQ-Bench Creative Writing | 1553 | — |
| EQ-Bench 4 | 1166 | — |
Frequently asked questions
Is DeepSeek V4 Pro better than Mercury 2?
DeepSeek V4 Pro is the stronger model overall, scoring 54.2 to 38.8 on the Noometry Index.
Which is cheaper, DeepSeek V4 Pro or Mercury 2?
Mercury 2 is cheaper. It lists at $0.25 per million input tokens and $0.75 per million output tokens; DeepSeek V4 Pro lists at $0.43 and $0.87.
Is DeepSeek V4 Pro or Mercury 2 better for coding?
DeepSeek V4 Pro scores higher on coding benchmarks: 52.2 versus 33.2 in the Noometry coding category.
Which has the bigger context window?
DeepSeek V4 Pro does, with 1M tokens against 128K.
How many benchmarks do DeepSeek V4 Pro and Mercury 2 share?
17 benchmarks have published results for both models. DeepSeek V4 Pro has 48 scored results on Noometry and Mercury 2 has 17.