Anthropic's best AI model struggles to attract users as cheaper tools thrive
TL;DR Highlight
Top-tier model reportedly loses users to cheaper options over price, limits, and missing ZDR.
Who Should Read
Developers choosing between Claude, GPT, and open-weight models for coding agents and API costs
Core Mechanics
- Article body was not provided
- Without the FT text, the article itself cannot be summarized
- The title says Anthropic's best model struggles to attract users while cheaper tools thrive
Community Reactions
- Opinion that Fable is best but hard to use
- One commenter said usage was capped below 50%, so one review took a month
- One commenter said Sol reached 70-90% of Fable on the same code review
- Experience running Fable for 18 hours autonomously
- A commenter moved a project from Spark to Pandas using Opus sub-agents
- It used 2B tokens over 24 hours, about $2000 at API prices
- Opinion that ZDR is the enterprise blocker
- One commenter's company could not roll Fable out widely
- Commenters said this rules it out for many businesses
- Reports of moving to local or open-weight models
- A commenter uses Kimi K3 as a daily driver
- Another said local models like Qwen 27B handle routine tasks at lower cost
How to Apply
- Compare two models on the same code review
- Run one repo diff through a premium and a cheaper model today
- Tabulate bugs found against cost
- Route tasks by difficulty
- Send only the hardest new features to the top model
- Send routine work to cheaper or local models
- Check ZDR support before adopting
- For services handling customer data, confirm each model's ZDR support first
Try It
# Paste into Claude or ChatGPT (simple version based on commenter experience; the source has no method)
You are a code reviewer. Review the diff below and list bugs, security issues, and improvements in a table sorted by severity.
{{diff}}
Next: run the same prompt on another model and compare findings and cost.Terminology
Related Papers
Mercury 3: Diffusion at scale
Scaling diffusion LLM 40x gives Mercury 3: 4–8x faster than speed-tier models, similar quality
Timestral: Mistral's time-series foundation models
A 209M time-series model ranks first on three benchmarks and runs on CPU.
DeepSeek V4.1 Flash 1017 (Max)
DeepSeek V4.1 Flash 1017 Max scores 45 on the Intelligence Index at 211 tok/s and $0.30 input.
If you don't run your own LLM proxy, how do you know whats going on?
Running your own LLM proxy shows which keys, models, and cache misses drive your usage.
DeepSeek: DeepSeek v4.1 Flash 1017
DeepSeek V4.1 Flash 1017: $0.05/M input, 1M context, built for agents
DeepSeek-v4.1-Flash-1017
DeepSeek V4.1 Flash minor update lifts coding and agent benchmarks while cutting run time and output length
Show HN: Llmspeed – feel how fast a local LLM runs before you buy the hardware