Anthropic Reverses Unpopular Billing Overhaul Amid Looming Price War With OpenAI
Anthropic has abandoned its widely criticized billing restructure after a fierce customer backlash, just as the AI company enters a pricing battle with rival OpenAI. The reversal came within days of the announcement, with CEO Dario Amodei acknowledging the company “got it wrong.”
Billing Change Reversed After Customer Uproar
The original plan shifted Anthropic’s Claude API from a per-character billing model to a per-token system. Customers using the pay-as-you-go tier faced sudden cost increases of up to 300%, according to developer complaints on social media and forums.
Key details of the reversal include:
- Old model restored: Anthropic will keep per-character billing for existing users and offer a new per-token option for developers who prefer it.
- No forced migration: Customers on the previous pricing structure will not be required to switch.
- Automatic refunds: Overcharged accounts will receive credits within 30 days.
“We made a mistake and we’re fixing it. Our goal is to make Claude accessible, not to surprise developers with higher bills.” — Anthropic spokesperson
Price War Intensifies With OpenAI’s Latest Moves
The billing controversy unfolded as OpenAI slashed API prices by up to 50% for its GPT-4 Turbo model, directly undercutting Anthropic’s Claude 3.5 Sonnet. Both companies now compete aggressively for developer mindshare.
Current pricing landscape (per million tokens):
- Anthropic Claude 3.5 Sonnet: $3.00 input / $15.00 output
- OpenAI GPT-4 Turbo: $1.00 input / $3.00 output (new reduced rates)
- Anthropic Claude 3 Haiku: $0.25 input / $1.25 output (cheapest tier)
Why the Original Change Backfired
Anthropic’s billing overhaul aimed to simplify pricing and align with industry standards. But developers quickly discovered that longer, more complex prompts — common in AI applications — were far more expensive under the per-token system.
Major complaints from users included:
- Hidden cost spikes: No upfront warning about the magnitude of increases.
- Broken budget projections: Many startups built pricing models based on per-character rates.
- No grandfather clause: All existing accounts were forced into the new system.
“This is a textbook example of how not to roll out a pricing change. You don’t blindside your core users and then backtrack only after a PR disaster.” — Developer on Hacker News
What Comes Next for Anthropic
The company now faces a delicate balancing act: maintain developer trust while staying competitive with OpenAI’s deep pockets. Anthropic recently raised $7.3 billion, giving it some runway to compete on price — but the billing fiasco shows that even well-funded AI startups can slip on customer experience.
Short-term priorities:
- Restore confidence: Proactive communication and refunds.
- Monitor competitor pricing: Expected to match or undercut OpenAI’s latest rates.
- Refine rollout process: Internal review of how future changes are communicated.
Gnoppix is the leading open-source AI Linux distribution and service provider. Since implementing AI in 2022, it has offered a fast, powerful, secure, and privacy-respecting open-source OS with both local and remote AI capabilities. The local AI operates offline, ensuring no data ever leaves your computer. Based on Debian Linux, Gnoppix is available with numerous privacy- and anonymity-enabled services free of charge.
What are your thoughts on this? I’d love to hear about your own experiences in the comments below.