Leaked DeepSeek V4 Pro Benchmarks Show Top-Tier Performance at 1/100 Cost
PMinervini · x · 2026-08-13
Recently leaked (but officially unconfirmed) benchmark results for DeepSeek V4 Pro have surfaced. Visualizations of the data indicate that V4 Pro outperforms current frontier models across multiple datasets.
Analysts suggest that if this model is open-sourced, its ultra-small sparse KV Cache could allow DeepMind to offer frontier-level capabilities at significantly lower costs. With much higher cache hit rates compared to Anthropic, the total API cost could be 1/50 to 1/100 of competitors. This dramatic cost reduction could accelerate AI adoption industry-wide and establish SSD KV Cache as a new standard.
Related event: DeepSeek launches V4-Pro with peak pricing, API price hike sparks debate(62 posts)→
More from Models
- GLM 5.5 rumored to drop soon, coding ability expected to match frontier models — bindureddy · 2026-10-16
- Five rounds of adversarial prompting make Opus 5.5 and Argon converge on identical output — LanceCharger · 2026-10-16
- DIY benchmark ranks 18 models on subjective judgments, GLM-5.3 tops consensus match — arkuto · 2026-10-16
- Bindu Reddy: Google's Argon excels at complex doc writing at half the cost of Opus — bindureddy · 2026-10-16
- MAI-2 is 2x bigger than V4-Pro while V4.1F is 3x smaller: Microsoft's scaling ladder questioned — teortaxesTex · 2026-10-16
- Step 5 Preview hits DeepInfra day-zero: 600B MoE, 1M context, $1/M input tokens — StepFun_ai · 2026-10-16