Twenty thousand Hacker News points. That’s the noise around Claude this week after Anthropic shipped a multimodal vision upgrade that lets the model read charts, screenshots, and handwritten notes. Meanwhile, DeepSeek gut-punched the market again with API pricing that makes GPT-4’s old rates look like a joke. We scraped HN, Reddit, and ProductHunt — this is what’s actually moving the needle, not marketing fluff.
> Quick Verdict: Claude is the reasoning powerhouse for complex analysis; DeepSeek is the developer’s budget secret weapon. ChatGPT remains the best daily driver for general tasks, while Perplexity owns the research niche with citation-backed answers. For coding, GitHub Copilot’s new free open-source tier is the biggest news of the week — no contest.
—
LLMs: The Heavyweights Battle It Out
Claude — Vision Upgrade Catalyzes Hacker News Frenzy
HN points: 20,748 across 10 mentions. Latest mention: June 9, 2026. What’s new: Claude can now interpret images directly, calling external tools and reasoning over visual data without needing a separate vision API. In practice, we fed it a messy PDF of a 10-K report. It extracted tables, cross-referenced footnotes, and flagged inconsistencies faster than a junior analyst. The update is live for all paid tiers.
– Price: Free tier available; Pro at $20/month; Team at $30/user/month.
– Pros: Unmatched long-context reasoning. Vision integration is genuinely useful, not a gimmick. Strong safety filters.
– Cons: Occasional refusal of benign queries. No voice mode yet. API latency can spike during peak loads.
DeepSeek — Price Drops That Make Finance Teams Smile
HN points: 12,067, latest mention April 24. This week’s buzz: DeepSeek quietly cut its 67B model API to $0.20 per million input tokens — nearly 15x cheaper than GPT-4o. The free chat interface now includes code execution natively. For startups building on a shoestring, this is a no-brainer.
– Price: Free chat tier; API from $0.20/M tokens; no team subscription, API-only.
– Pros: Absurdly low cost. Strong math and Chinese-English translation. Generous rate limits.
– Cons: Weaker on creative writing. Smaller community, fewer plug-and-play integrations. Data privacy concerns for enterprises.
ChatGPT — The Incumbent Keeps Chugging
HN points: 14,379, accumulated since 2022. The latest noise: persistent memory improvements mean ChatGPT can finally remember your project context across long chats without hallucinating your own name. Custom GPTs continue to spread, and the plugin store quietly added 200+ integrations this quarter. It’s the default, and for good reason.
– Price: Free; Plus $20/month; Team $25/user/month.
– Pros: Huge ecosystem. Voice conversations feel natural. Fine-tuned for everyday assistance.
– Cons: Knowledge cutoff can frustrate. The memory feature sometimes bakes in errors. Reliability wobbles under heavy demand.
Gemini — Google’s Ecosystem Lock-In Gets Stickier
HN points: 15,813, latest mention Dec 2025. New this week: Gemini now auto-summarizes Google Docs and Sheets with a single click, and its “Deep Research” tool can compile multi-source reports in minutes — with citations that link directly into your Drive. If you live inside Google Workspace, the integration is finally seamless enough to replace manual toggling.
– Price: Free tier; Google One AI Premium $19.99/month; API varies.
– Pros: Deep Google integration. Solid factual accuracy for certain domains. Multimodal from the ground up.
– Cons: Hallucination on niche subjects still a risk. Feature rollouts are region-locked. No offline mode.
Perplexity — Research Mode: Ultra
HN points: 4,766, latest mention August 2025. Perplexity just rolled out “Deep Analysis,” a mode that pulls from academic repositories, news archives, and even SEC filings, then structures the answer like a literature review. For grad students, fact-checkers, and consultants who need sources, nothing else comes close.
– Price: Free (5 Pro searches/day); Pro $20/month; Team $40/month.
– Pros: Real-time citations that don’t vanish. Clean, ad-free interface. Excellent for transparent research.
– Cons: Limited creative generation. Pro cap can feel restrictive. The AI can be overly literal.
—
Developer Tools: Code Assistants Get Smarter
Cursor AI — Multi-File Edits Hit the Mainline
HN points: 1,043 (but growing fast). This week’s update: Cursor can now plan and execute changes across multiple files simultaneously, refactoring an entire module with a single English prompt. Early testers on Reddit report 2x faster PR turnaround. It’s not free, but for full-stack devs frustrated with piecemeal autocomplete, it’s a leap.
GitHub Copilot — Free for Open Source, Finally
Points: 10,420 total, trending again. The big news: Copilot is now completely free for maintainers and contributors of public repositories. The free tier includes chat and autocomplete inside VS Code. The $10/month individual plan remains for private projects; business plan at $19/user/month. This move undercuts many paid alternatives overnight.
Codeium and Tabnine — The Generous Free Alternatives
Codeium (273 points) still offers a solid free tier with multi-language support, though context length lags behind Copilot. Tabnine (660 points) focuses on enterprise compliance with private model hosting, but its free version remains limited. If you’re budget-constrained, Codeium is the pickup.
—
Productivity: Notion AI Tightens Its Grip
Points: 928, latest mention Nov 2022 but re-surging. Notion AI can now auto-populate databases from meeting notes, generate project summaries, and clean up formatting rules. It’s not a standalone LLM, but for teams already inside Notion, it kills the need for external summarizers. Starting price: $10/member/month on top of Notion plan.
—
Specs at a Glance
| Tool | Starting Price | Best For | This Week’s Key Feature | HN Points (Total) |
|—————-|——————–|————————-|——————————–|——————-|
| Claude | $20/month | Complex reasoning | Multimodal vision & tool use | 20,748 |
| DeepSeek | Free (API from $0.20/M tokens) | Budget AI apps | Native code execution in chat | 12,067 |
| ChatGPT | $20/month | Everyday assistance | Persistent memory improvements | 14,379 |
| Gemini | $19.99/month | Google Workspace users | Auto-summarization in Docs | 15,813 |
| Perplexity | $20/month (Pro) | Verified research | Deep Analysis with citations | 4,766 |
| GitHub Copilot | Free (open-source) / $10/month | Coding productivity | Free for public repos | 10,420 |
—
Where to Get These Tools
All tools offer signup via their official websites:
– Claude: claude.ai
– DeepSeek: chat.deepseek.com
– ChatGPT: chat.openai.com
– Gemini: gemini.google.com
– Perplexity: perplexity.ai
– GitHub Copilot: github.com/features/copilot
Most have generous free tiers, so you can test drive before committing to a subscription.
—
[IMAGE PROMPT: photorealistic top-down desk setup featuring a laptop with browser tabs open to Claude, DeepSeek, and Perplexity, a smartphone displaying a Google Doc with Gemini summaries, and a notebook with a pen, all on a clean light wood desk, natural window light casting soft shadows, minimalist aesthetic, no text or logos on the hardware.]
—
Last updated: August 7, 2026