kimi-k2.7-code vs claude-fable-5

Pricing, Performance & Features Comparison

kimi-k2.7-code

Authormoonshot

Context Length262K

Reasoning

Providers1

ReleasedJun 2026

Knowledge Cutoff–

License–

Moonshot’s coding- and agent-specialized model built on Kimi K2.6; thinking is always on.

Input$0.95

Output$4

Latency (p50)5.4s

Output Limit–

Function Calling

JSON Mode

InputText

OutputText

moonshot

in$0.95out$4cache$0.19–

Latency (24h)

Success Rate (24h)

claude-fable-5

Authoranthropic

Context Length1M

Reasoning

Providers1

ReleasedJun 2026

Knowledge Cutoff–

License–

Anthropic’s most capable widely released model, a tier above Opus. Thinking is always on, sampling parameters are rejected, and requests may return a refusal stop reason.

Input$10

Output$50

Latency (p50)8.2s

Output Limit128K

Function Calling

JSON Mode

InputText, Image

OutputText

anthropic

in$10out$50cache$1write$13