model comparison

vs

Claude Opus 5 vs Kimi K3

Anthropic against Moonshot AI, compared on listed price, real workload cost, context and capabilities.

the short answer

  • Kimi K3 lists lower input and output prices.
  • Kimi K3 reads cached input for less ($0.3/M against Claude Opus 5's $0.5/M), which decides the bill for agents that resend long context.
  • Claude Opus 5 retired on 4 October 2026; its numbers are kept for reference.

side-by-side specs

specClaude Opus 5Kimi K3
providerAnthropicMoonshot AI
input price$5/M$3/M
output price$25/M$15/M
cached input$0.5/M$0.3/M
blended price (3:1)$10/M$6/M
context window1M tokens1M tokens
max output128k tokens—
inputtext, imagestext, images
structured outputyesyes
weightsclosedopen (Kimi K3 License)
releasedJuly 24, 2026July 2026
availabilityretired 4 October 2026live

same work, both models

workloadClaude Opus 5Kimi K3
one chat reply3,000 tokens in, 500 outClaude Opus 5 costs 1.7× as much as Kimi K3$0.03$0.02
summarise a 100-page report50,000 tokens in, 1,500 out, one request$0.29$0.17
1,000 classification calls500 tokens in, 20 out, per call$3.00$1.80
coding-agent session2M tokens in across ~60 turns, 60,000 out, no cache$11.50$6.90
same session, cached85% of input read from the prompt cache, the rest written to itClaude Opus 5 costs 1.8× as much as Kimi K3$4.23$2.31

at listed rates. token counts are typical sizes, not measurements.

Claude Opus 5 pricing notes

  • cache writes bill at the 5-minute rate ($6.25 per 1M); Anthropic's 1-hour tier ($10) is not modelled.
  • Anthropic's fast mode costs twice the base price; the listed rate is standard speed.

Kimi K3 pricing notes

  • not included in any plan: every request bills from credits.
  • context caching is automatic: cache hits bill at $0.30 per 1M input tokens, a tenth of the input rate.
  • thinking is on by default at max effort, and reasoning tokens bill at the output rate.

when to pick which

Claude Opus 5

Claude Opus 5 was Anthropic's Opus model released July 2026, at Opus 4.8's $5 / $25 price, with adaptive thinking and text plus image input. it was retired in favour of Claude Opus 5.5, which costs less at $4 / $20.

pick it if

  • you wanted agentic coding and computer use Anthropic placed near Fable 5
  • you wanted Fable 5-class work at half of Fable 5's price
  • you needed image input alongside adaptive thinking

skip it if

  • you are choosing today; Opus 5.5 replaced it at a lower $4 / $20

Kimi K3

Moonshot AI's most capable Kimi model: an open-weights 2.8T-parameter mixture-of-experts with native vision and a 1M-token context. it sits above Kimi 2.6 and Kimi 2.7 Code, which have a 256K context and cost less per token.

pick it if

  • you want the model Moonshot calls its most capable
  • your prompts outgrow the 256K context of Kimi 2.6 and 2.7 Code
  • you resend long prompts, so cache hits bill at a tenth of input

skip it if

  • Kimi 2.6 or 2.7 Code covers your workload at a lower rate

faq

which is cheaper, Claude Opus 5 or Kimi K3?

Kimi K3 lists lower input and output prices. on a coding-agent session (2M tokens in, 60,000 out) Claude Opus 5 costs $11.50 and Kimi K3 $6.90.

try either model

switch between models inside the same octomind session.

prices from the catalog of 2026-11-18