# DeepSeek V4 Flash 0731 As of October 22, 2026, DeepSeek V4 Flash 0731 ranks #54 of 123 models overall with 55.2 ±8.0 on the LLMBenchmark Index. DeepSeek V4 Flash 0731 ranks #46 of 122 in reasoning, its strongest category. DeepSeek V4 Flash 0731 costs $0.14 per million input tokens and $0.396 per million output tokens. DeepSeek V4 Flash 0731 was released on July 31, 2026. Page: https://llmbenchmark.ai/models/deepseek-v4-flash-0731 Data from October 22, 2026. Every figure below names the board it was read from and the day. - Maker: DeepSeek - Released: July 31, 2026 - Price per million input tokens: $0.14 - Price per million output tokens: $0.396 - Context window: 1.05M tokens - Overall rank: 54 of 123, by the LLMBenchmark Index - LLMBenchmark Index: 55.2 ±8.0 (90% margin, from 6 of 7 inputs) - Artificial Analysis Intelligence Index: 34.3 - Vibe Coding Index: 51.8 (from 4 of 6 coding benchmarks) Benchmark results (55 of 195): - GPQA Diamond: 89.9% (source: GPQA (Vals AI), read October 22, 2026) - MMLU Pro: 86.2% (source: MMLU Pro (Vals AI), read October 22, 2026) - AA Intelligence Index: 34.3 (source: Artificial Analysis (via OpenRouter), read October 22, 2026) - AA Coding Index: 69.1 (source: Artificial Analysis (via OpenRouter), read October 22, 2026) - AA Agentic Index: 41.0 (source: Artificial Analysis (via OpenRouter), read October 22, 2026) - Design Arena: websites: 1237 Elo (source: Design Arena, read October 22, 2026) - Design Arena: UI components: 1233 Elo (source: Design Arena, read October 22, 2026) - Design Arena: game dev: 1207 Elo (source: Design Arena, read October 22, 2026) - Design Arena: data viz: 1180 Elo (source: Design Arena, read October 22, 2026) - Design Arena: 3D: 1208 Elo (source: Design Arena, read October 22, 2026) - Design Arena: all code categories: 1229 Elo (source: Design Arena, read October 22, 2026) - Design Arena: SVG: 1184 Elo (source: Design Arena, read October 22, 2026) - Design Arena: ASCII art: 1090 Elo (source: Design Arena, read October 22, 2026) - LiveCodeBench (Vals AI): 87.3% (source: Vals AI boards, read October 22, 2026) - SWE-bench Verified (Vals AI): 88.8% (source: Vals AI boards, read October 22, 2026) - Terminal-Bench 4 (Vals AI): 18.7% (source: Vals AI boards, read October 22, 2026) - Vibe Code Bench (Vals AI): 74.7% (source: Vals AI boards, read October 22, 2026) - Code migration (Vals AI): 38.6% (source: Vals AI boards, read October 22, 2026) - ProgramBench (Vals AI): 0.0% (source: Vals AI boards, read October 22, 2026) - IOI (Vals AI): 32.7% (source: Vals AI boards, read October 22, 2026) - Terminal-Bench 2.1 (Vals AI): 67.0% (source: Vals AI boards, read October 22, 2026) - SkillsBench (Vals AI): 50.7% (source: Vals AI boards, read October 22, 2026) - CyberBench (Vals AI): 72.1% (source: Vals AI boards, read October 22, 2026) - Finance Agent (Vals AI): 49.5% (source: Vals AI boards, read October 22, 2026) - Tax Agent Bench (Vals AI): 28.1% (source: Vals AI boards, read October 22, 2026) - Legal Research Bench (Vals AI): 30.3% (source: Vals AI boards, read October 22, 2026) - Excel Modeling Benchmark (Vals AI): 57.0% (source: Vals AI boards, read October 22, 2026) - BioMysteryBench (Vals AI): 64.4% (source: Vals AI boards, read October 22, 2026) - Terminal-Bench-Science (Vals AI): 4.3% (source: Vals AI boards, read October 22, 2026) - LegalBench (Vals AI): 77.7% (source: Vals AI boards, read October 22, 2026) - CorpFin (Vals AI): 61.9% (source: Vals AI boards, read October 22, 2026) - TaxEval (Vals AI): 70.7% (source: Vals AI boards, read October 22, 2026) - MedCode (Vals AI): 41.4% (source: Vals AI boards, read October 22, 2026) - MedScribe (Vals AI): 80.4% (source: Vals AI boards, read October 22, 2026) - GPQA Diamond (Artificial Analysis run): 90.8% (source: Artificial Analysis, read October 22, 2026) - Humanity's Last Exam (Artificial Analysis run): 38.6% (source: Artificial Analysis, read October 22, 2026) - SciCode: 50.3% (source: Artificial Analysis, read October 22, 2026) - Long context reasoning (AA-LCR): 79.7% (source: Artificial Analysis, read October 22, 2026) - FrontierCode: 18.8% (source: Epoch AI benchmarking hub, read October 22, 2026) - WeirdML: 63.0% (source: Epoch AI benchmarking hub, read October 22, 2026) - GPQA Diamond (Epoch AI run): 91.0% (source: Epoch AI benchmarking hub, read October 22, 2026) - ARC-AGI: 89.0% (source: Epoch AI benchmarking hub, read October 22, 2026) - ARC-AGI-2: 61.4% (source: Epoch AI benchmarking hub, read October 22, 2026) - SimpleBench: 61.1% (source: Epoch AI benchmarking hub, read October 22, 2026) - Chess puzzles: 32.0% (source: Epoch AI benchmarking hub, read October 22, 2026) - Mystery game puzzles: 34.0% (source: Epoch AI benchmarking hub, read October 22, 2026) - OTIS Mock AIME: 94.4% (source: Epoch AI benchmarking hub, read October 22, 2026) - FrontierMath, tiers 1 to 3: 57.5% (source: Epoch AI benchmarking hub, read October 22, 2026) - FrontierMath, tier 4: 24.4% (source: Epoch AI benchmarking hub, read October 22, 2026) - ProofBench: 56.0% (source: Epoch AI benchmarking hub, read October 22, 2026) - SimpleQA Verified: 33.6% (source: Epoch AI benchmarking hub, read October 22, 2026) - DTBench: 90.9% (source: Epoch AI benchmarking hub, read October 22, 2026) - LMCA: 41.7% (source: Epoch AI benchmarking hub, read October 22, 2026) - ALE-Bench: 1306.1 (source: ALE-Bench (Sakana AI), read October 22, 2026) - Epoch Capabilities Index: 154.8 (source: Epoch AI benchmarking hub, read October 22, 2026)