Model comparison
DeepSeek V4 Pro vs Hy4 preview
DeepSeek V4 Pro is the stronger model overall, scoring 54.3 to 45.3 on the Noometry Index.
Last verified . 3 shared benchmarks.
DeepSeek
TencentSummary
- They share 3 benchmarks with published results for both. DeepSeek V4 Pro scores higher in 3 categories and Hy4 preview in 0 categories; 2 gaps are clear of the uncertainty.
- The widest gap is in reasoning, where DeepSeek V4 Pro leads 56.5 to 31.9.
- The biggest single-benchmark swing is ProofBench: 50% for DeepSeek V4 Pro and 75% for Hy4 preview.
- DeepSeek V4 Pro is cheaper at $0.66 / $1.98 per million input/output tokens, against $0.75 / $2.25 for Hy4 preview.
- Hy4 preview accepts more context: 1.05M tokens versus 1M.
Side by side
| DeepSeek V4 Pro | Hy4 preview | |
|---|---|---|
| Provider | DeepSeek | Tencent |
| Noometry Index | 54.3 | 45.3 |
| Released | 2026-04-24 | 2026-08-28 |
| Weights | Open | Open |
| Context window | 1M | 1.05M |
| Max output | 393K | 64K |
| Input $ / M tokens | $0.66 | $0.75 |
| Output $ / M tokens | $1.98 | $2.25 |
| Results tracked | 48 | 3 |
Sponsored placements are available on pages like this one. Advertise on Noometry
Category by category
Coding Too close to call
DeepSeek V4 Pro: 52.4 (#34), Hy4 preview: 51.6 (#38)
| Benchmark | DeepSeek V4 Pro | Hy4 preview |
|---|---|---|
| LMArena WebDev | 1582 | 1632 |
| SWE-bench Verified | 77.6% | — |
| FrontierCode | 28.6% | — |
| SciCode | 51% | — |
| WeirdML | 66.2% | — |
| LMArena Coding | 1470 | — |
| ALE-Bench | 1,403 | — |
Agentic & Tool Use Not comparable
DeepSeek V4 Pro: 32.8 (#58), Hy4 preview: —
| Benchmark | DeepSeek V4 Pro | Hy4 preview |
|---|---|---|
| APEX-Agents | 47.3% | — |
| Vending-Bench 2 | 3,285 | — |
Reasoning DeepSeek V4 Pro leads
DeepSeek V4 Pro: 56.5 (#24), Hy4 preview: 31.9 (#79)
| Benchmark | DeepSeek V4 Pro | Hy4 preview |
|---|---|---|
| NYT Connections (extended) | 91.3% | 68.2% |
| ARC-AGI-2 | 61.3% | — |
| Kagi LLM Benchmark | 53.5% | — |
| ARC-AGI-1 | 90.5% | — |
| CritPt | 18% | — |
| Chess Puzzles | 47% | — |
| LMArena Hard Prompts | 1461 | — |
| Mystery Game Puzzles | 43% | — |
| DTBench | 93.9% | — |
| LMCA | 45.5% | — |
| Surface Evolver Bench | 40% | — |
| Epoch Capabilities Index | 155.31 | — |
| ForecastBench | 56.1 | — |
Math DeepSeek V4 Pro leads
DeepSeek V4 Pro: 64.8 (#30), Hy4 preview: 55.7 (#42)
| Benchmark | DeepSeek V4 Pro | Hy4 preview |
|---|---|---|
| ProofBench | 50% | 75% |
| FrontierMath (Tiers 1-3) | 64.6% | — |
| FrontierMath Tier 4 | 26.8% | — |
| MathArena Final-Answer Competitions | 76.6% | — |
| OTIS Mock AIME 2024-2025 | 98.6% | — |
| LMArena Math | 1455 | — |
Knowledge Not comparable
DeepSeek V4 Pro: 59.5 (#31), Hy4 preview: —
| Benchmark | DeepSeek V4 Pro | Hy4 preview |
|---|---|---|
| GPQA Diamond | 91.7% | — |
| SimpleQA Verified | 52.9% | — |
| Vectara Hallucination Rate | 8.6% | — |
| LMArena Expert | 1464 | — |
Multilingual Not comparable
DeepSeek V4 Pro: 54.4 (#45), Hy4 preview: —
| Benchmark | DeepSeek V4 Pro | Hy4 preview |
|---|---|---|
| LMArena Non-English | 1439 | — |
| LMArena Chinese | 1486 | — |
| LMArena French | 1472 | — |
| LMArena German | 1458 | — |
| LMArena Japanese | 1445 | — |
| LMArena Korean | 1447 | — |
| LMArena Russian | 1453 | — |
| LMArena Spanish | 1458 | — |
Instruction Following Not comparable
DeepSeek V4 Pro: 76.1 (#47), Hy4 preview: —
| Benchmark | DeepSeek V4 Pro | Hy4 preview |
|---|---|---|
| LMArena Instruction Following | 1448 | — |
Long Context Not comparable
DeepSeek V4 Pro: 45.0 (#51), Hy4 preview: —
| Benchmark | DeepSeek V4 Pro | Hy4 preview |
|---|---|---|
| CL-bench Life | 13.5% | — |
| LMArena Longer Query | 1458 | — |
Writing & Preference Not comparable
DeepSeek V4 Pro: 65.5 (#46), Hy4 preview: —
| Benchmark | DeepSeek V4 Pro | Hy4 preview |
|---|---|---|
| LMArena Text | 1451 | — |
| LMArena Creative Writing | 1446 | — |
| EQ-Bench Creative Writing | 1553 | — |
| EQ-Bench 4 | 1166 | — |
| LMArena Multi-Turn | 1467 | — |
Frequently asked questions
Is DeepSeek V4 Pro better than Hy4 preview?
DeepSeek V4 Pro is the stronger model overall, scoring 54.3 to 45.3 on the Noometry Index.
Which is cheaper, DeepSeek V4 Pro or Hy4 preview?
DeepSeek V4 Pro is cheaper. It lists at $0.66 per million input tokens and $1.98 per million output tokens; Hy4 preview lists at $0.75 and $2.25.
Is DeepSeek V4 Pro or Hy4 preview better for coding?
They score almost the same on coding (52.4 vs 51.6); test both on your own repository before choosing.
Which has the bigger context window?
Hy4 preview does, with 1.05M tokens against 1M.
How many benchmarks do DeepSeek V4 Pro and Hy4 preview share?
3 benchmarks have published results for both models. DeepSeek V4 Pro has 48 scored results on Noometry and Hy4 preview has 3.