I've been running Kimi K3 alongside Claude on normal coding work and can't tell them apart—same tasks, quality, and token counts. Yet pricing is nowhere near each other: K3's API runs $3 per million input tokens and $15 per million output versus Claude's $10 and $50. Subscriptions are even more lopsided. Kimi's paid plans start at $19 a month, with a $39 coding tier far more generous than Claude's tightly metered allowances that can exhaust before lunch.

Claude couldn't sustain Fable access on the twenty dollar plan, so they turned it off and the plan quietly falls back to Opus. When the headline model can be switched off because economics don't work, the plan was never really selling you that model. Kimi's tiers don't come with that asterisk.

The bigger story is what an unmitigated failure US AI policy has been. The administration held Fable back, shipping a hindered version that refuses whole categories of work. Meanwhile a frontier-quality model with none of those restrictions is a download away, released by a Chinese lab the US government cannot regulate. Semgrep found GLM 5.2 beating Claude on cyber benchmarks for exactly this reason—the restricted model declines the work and the open one just does it.

I expect the government to reach for the auto-industry playbook: decades of subsidies and protective tariffs producing domestic models that only get used inside the US and can't compete internationally. Until then, I can't come up with a reason to keep paying for Claude.