First screen, captured 17 hours ago. Hover to scroll the page, or open the video.

OpenAI and Anthropic cut their LLM prices by up to 80% within a few months. Why?

What HN thinks

Prices fell 80% and the thread immediately split into 'competition', 'bubble', and 'ban the Chinese models' camps, with a detour into US party-system numbering.

Cheap open-weight competition
Chinese open models like Kimi, GLM, DeepSeek and Qwen are good enough for most tasks at far lower prices, forcing the US labs to cut. Some expect the labs to lobby for regulation or bans to protect themselves.
Labs are financially unsustainable
Price cuts on top of huge capex and slowing revenue growth suggest a race to the bottom, rushed IPOs, possible bailouts, or an overbuilt GPU bubble like the telecom crash.
Inference margins were always fat
DeepSeek's published ~$38 per GPU-hour figures suggest inference is very profitable, so the labs had plenty of room to cut; the open question is training and R&D costs.

Consensus: Most agree competition, especially from cheaper open-weight models, is a major driver, but there is no agreement on whether the frontier labs can stay profitable.

Recommended in the comments

OpenCode (tool): Several use it with open models like DeepSeek and Kimi to avoid lock-in.
Zed (product): One commenter found its agent gets better results than OpenCode from the same model.
Crush ↗ (tool): Recommended as great and easy to hack on.
OpenRouter (service): Cheap API access that makes buying local inference hardware hard to justify.
Claude Code (tool): One commenter found it the best harness, usable with any compatible API.

Summarised by Claude from the comment thread 5 hours ago.

Points: 365
Comments: 347
Best rank: #8
Posted: 2026-10-25 by minimaxir
Category: Ai ml

llm-pricing openai anthropic open-weight-models deepseek openrouter ai-economics

All captures:

Full page

Captured 17 hours ago. Click for the whole screenshot.

Looks similar

Screenshot of minimaxir.com: The mysterious Hy3 LLM is topping OpenRouter Model Rankings by a large margin

The mysterious Hy3 LLM is topping OpenRouter Model Rankings by a large margin (minimaxir.com)

131 points by freediver 5 months ago | 103 comments

Screenshot of simonwillison.net: I think Anthropic and OpenAI have found product-market fit

I think Anthropic and OpenAI have found product-market fit (simonwillison.net)

1048 points by simonw 5 months ago | 1153 comments

Simon Willison blog post arguing Anthropic and OpenAI have reached product-market fit as enterprises pay API prices for LLM usage and coding agents.

anthropic openai product-market-fit llm-pricing

Screenshot of blog.doubleword.ai: The gap between open weights LLMs and closed source LLMs

The gap between open weights LLMs and closed source LLMs (blog.doubleword.ai)

285 points by kkm 4 months ago | 213 comments

Screenshot of artificialanalysis.ai: GLM-5.2 is the new leading open weights model on Artificial Analysis

GLM-5.2 is the new leading open weights model on Artificial Analysis (artificialanalysis.ai)

845 points by himata4113 4 months ago | 411 comments

Artificial Analysis article reporting that Z ai's GLM-5.2 scores 51 on its Intelligence Index, leading open weights models ahead of MiniMax-M3 and DeepSeek V4 Pro.

glm-5-2 zhipu-ai open-weights-models llm-benchmarks

Screenshot of seangoedecke.com: Sycophancy is the first LLM "dark pattern"

Sycophancy is the first LLM "dark pattern"(seangoedecke.com)

135 points by jxmorris12 10 months ago | 81 comments

Screenshot of minimaxir.com: Can LLMs write better code if you keep asking them to “write better code”?

Can LLMs write better code if you keep asking them to “write better code”?(minimaxir.com)

710 points by rcarmo 1 year ago | 411 comments

Max Woolf's blog post experimenting with whether repeatedly prompting an LLM to "write better code" actually improves the code it generates.

llm code-generation prompt-engineering claude

Full page

Looks similar

The mysterious Hy3 LLM is topping OpenRouter Model Rankings by a large margin (minimaxir.com)

I think Anthropic and OpenAI have found product-market fit (simonwillison.net)

The gap between open weights LLMs and closed source LLMs (blog.doubleword.ai)

GLM-5.2 is the new leading open weights model on Artificial Analysis (artificialanalysis.ai)

Sycophancy is the first LLM "dark pattern"(seangoedecke.com)

Can LLMs write better code if you keep asking them to “write better code”?(minimaxir.com)

Related stories

OpenAI is too cheap to beat (generatingconversation.substack.com)

OpenAI, Google and Anthropic are struggling to build more advanced AI (bloomberg.com)

OpenAI dropped the price of o3 by 80%(twitter.com)

Why Math Dumped OpenAI (raymyers.org)

OpenAI researchers warned board of AI breakthrough ahead of CEO ouster (reuters.com)

OpenAI to become for-profit company (reuters.com)