DeepSeek V4 Ecosystem: From Open Weights to DeepClaude

Tracking the ripple effects of DeepSeek V4's open-weight release: near-frontier performance at a fraction of the cost, and how the community is already repurposing it in tools like DeepClaude.

Share: Long

LLM Hacker News 2d ago 1 min read

DeepSeek V4: Near-Frontier LLM Performance at a Fraction of the Cost

DeepSeek released DeepSeek-V4-Pro (1.6T total parameters, 49B active) and V4-Flash (284B total, 13B active), both Mixture-of-Experts models with MIT license and 1M token context. V4-Pro is the largest open-weights model released so far, and its pricing at $1.74/M input undercuts GPT-5.4 and Claude Sonnet 4.6 by more than half.

#deepseek #llm #open-weights

LLM Hacker News 12h ago 1 min read

DeepClaude: Run Claude Code's Agent Loop with DeepSeek V4 Pro at 17x Less Cost

DeepClaude keeps Claude Code's complete agent loop — file editing, bash, subagent spawning — while routing API calls to DeepSeek V4 Pro or other backends, cutting output token costs from $15/M to $0.87/M.

#claude-code #deepseek #developer-tools