Comparing AI Subscriptions: DeepSeek API vs. Claude Pro vs. Local LLMs
Unlikely_Bluejay5392 · reddit · 2026-08-24
A user detailed their journey testing various AI options, starting from free ChatGPT and Gemini Pro to paid subscriptions and local deployment.
Key Comparisons:
- Gemini: Weak coding skills, prone to agreeing with incorrect user corrections.
- Claude Pro ($20): Excellent speed and performance, but has a 5-hour usage limit.
- DeepSeek API: Great value using cache with Claude Code as an agent; however, OpenClaw drained funds rapidly.
- Ollama Local: Running 12-14B models (Q4) on an RTX 3060 12GB is fine for small fixes but too slow for complex coding projects.
Conclusion: After testing Cursor and Ollama Cloud, the user prefers returning to Claude Pro for top-tier model access.
More from coding & agent
- Developer habits shift: Agents become collaborators from simple tools — latticecut · 2026-08-24
- Dev bottleneck shifts from writing to reading code: exe.dev co-founder — thursdai_pod · 2026-08-24
- The biggest AI mistake: trying to reinvent the wheel instead of using tools — Tired40s · 2026-08-24
- DeepPaperNote turns research papers into Obsidian notes — tom_doerr · 2026-08-24
- Open Source RobotSoul: Persistent Identity for Agents After Context Resets — robauto-dot-ai · 2026-08-24
- Hermes Agent Masterclass Module 5: Providers & Local Models — NousResearch · 2026-08-24