# Claude Mythos Preview vs DeepSeek V4 Pro

> DeepSeek V4 Pro has enough public results to be ranked (#28); Claude Mythos Preview does not yet, so treat this comparison as directional.

- Canonical page: https://noometry.com/compare/claude-mythos-preview-vs-deepseek-v4-pro
- Last updated: 2026-10-10
- Shared benchmarks: 0

## Summary

- The widest gap is in agentic & tool use, where Claude Mythos Preview leads 41.1 to 32.7.
- DeepSeek V4 Pro has downloadable open weights; the other is API-only.

## Snapshot

| | Claude Mythos Preview | DeepSeek V4 Pro |
|---|---|---|
| Provider | Anthropic | DeepSeek |
| Noometry Index | 40.8 | 54.2 |
| Rank | — | 28 |
| Context | — | 1M |
| Input $/M | — | $0.43 |
| Output $/M | — | $0.87 |
| Weights | Proprietary | Open |

## Coding

- Claude Mythos Preview: —
- DeepSeek V4 Pro: 52.2 (#31)

| Benchmark | Claude Mythos Preview | DeepSeek V4 Pro |
|---|---|---|
| SWE-bench Verified | — | 77.6% |
| FrontierCode | — | 28.6% |
| LMArena WebDev | — | 1582 |
| SciCode | — | 51% |
| WeirdML | — | 66.2% |
| LMArena Coding | — | 1470 |
| ALE-Bench | — | 1,403 |

## Agentic & Tool Use

- Claude Mythos Preview: 41.1
- DeepSeek V4 Pro: 32.7 (#54)

| Benchmark | Claude Mythos Preview | DeepSeek V4 Pro |
|---|---|---|
| APEX-Agents | — | 47.3% |
| ExploitBench | 73.8% | — |
| METR Time Horizons | 85.2% | — |
| Vending-Bench 2 | — | 3,285 |

## Reasoning

- Claude Mythos Preview: —
- DeepSeek V4 Pro: 56.4 (#22)

| Benchmark | Claude Mythos Preview | DeepSeek V4 Pro |
|---|---|---|
| ARC-AGI-2 | — | 61.3% |
| Kagi LLM Benchmark | — | 53.5% |
| NYT Connections (extended) | — | 91.3% |
| 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 | — | 155.39 |
| ForecastBench | — | 56.1 |

## Math

- Claude Mythos Preview: —
- DeepSeek V4 Pro: 64.6 (#24)

| Benchmark | Claude Mythos Preview | DeepSeek V4 Pro |
|---|---|---|
| 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

- Claude Mythos Preview: —
- DeepSeek V4 Pro: 59.5 (#28)

| Benchmark | Claude Mythos Preview | DeepSeek V4 Pro |
|---|---|---|
| GPQA Diamond | — | 91.7% |
| SimpleQA Verified | — | 52.9% |
| Vectara Hallucination Rate | — | 8.6% |
| LMArena Expert | — | 1463 |

## Multilingual

- Claude Mythos Preview: —
- DeepSeek V4 Pro: 54.5 (#44)

| Benchmark | Claude Mythos Preview | DeepSeek V4 Pro |
|---|---|---|
| LMArena Non-English | — | 1438 |
| LMArena Chinese | — | 1486 |
| LMArena French | — | 1473 |
| LMArena German | — | 1458 |
| LMArena Japanese | — | 1448 |
| LMArena Korean | — | 1452 |
| LMArena Russian | — | 1452 |
| LMArena Spanish | — | 1457 |

## Instruction Following

- Claude Mythos Preview: —
- DeepSeek V4 Pro: 76.3 (#44)

| Benchmark | Claude Mythos Preview | DeepSeek V4 Pro |
|---|---|---|
| LMArena Instruction Following | — | 1448 |

## Long Context

- Claude Mythos Preview: —
- DeepSeek V4 Pro: 45.0 (#49)

| Benchmark | Claude Mythos Preview | DeepSeek V4 Pro |
|---|---|---|
| CL-bench Life | — | 13.5% |
| LMArena Longer Query | — | 1459 |

## Writing & Preference

- Claude Mythos Preview: —
- DeepSeek V4 Pro: 65.6 (#45)

| Benchmark | Claude Mythos Preview | DeepSeek V4 Pro |
|---|---|---|
| LMArena Text | — | 1451 |
| LMArena Creative Writing | — | 1446 |
| EQ-Bench Creative Writing | — | 1553 |
| EQ-Bench 4 | — | 1166 |
| LMArena Multi-Turn | — | 1467 |

## FAQ

### Is Claude Mythos Preview better than DeepSeek V4 Pro?

DeepSeek V4 Pro has enough public results to be ranked (#28); Claude Mythos Preview does not yet, so treat this comparison as directional.

### How many benchmarks do Claude Mythos Preview and DeepSeek V4 Pro share?

0 benchmarks have published results for both models. Claude Mythos Preview has 2 scored results on Noometry and DeepSeek V4 Pro has 48.