Kimi K2 0905 vs Kimi K2 Thinking.
model vs model · prices as of
Kimi K2 0905 takes 1 of 5 rounds against Kimi K2 Thinking. It is also the cheaper of the two to run, at $3 per million tokens against $3.41.
Moonshot AIwins 1 roundsKimi K2 0905
$3/M tokens
cheapest we can explain, at NanoGPT · $0.348 per 1,000 pages
262k context · tool calling · open weights
Moonshot AIKimi K2 Thinking
$3.41/M tokens
cheapest we can explain, at Vercel AI Gateway · $0.402 per 1,000 pages
262k context · tool calling · open weights
List price in (lower wins)
$0.60/M
$0.60/M
List price out (lower wins)
$2.50/M
$2.50/M
Cheapest we can explain (lower wins)
$3/M at NanoGPT
$3.41/M at Vercel AI Gateway
Standard price (lower wins)
$4.30/M
$4.30/M
Cost per 1,000 pages (lower wins)
$0.348
$0.402
Context
262k tokens
262k tokens
Open weights
yes
yes
Tool calling
yes
yes
Structured output
yes
yes
Reads images
no
no
Reasoning
no
yes
Released
4 Sept 2025
6 Nov 2025
Providers selling it
27
26
Price we measure against
the maker's own
the maker's own
Trains on your prompts (no is better)
not stated
not stated
Providers selling both · 17
Helicone$3.50 / $3.44
NanoGPT$3 / $4.30
OpenCode Zen$3.70 / $3.70
Vercel AI Gateway$4.01 / $3.41
TrustedRouter$4.54 / $3.17
io.net Intelligence$4.01 / $4.30
LLM Gateway$4.01 / $4.30
302.AI$4.43 / $4.03
Kimi K2 0905 then Kimi K2 Thinking, per million tokens · showing the 8 cheapest of 17
Go further
Add a third or fourth model, or switch the weighting, in the tool.
What moved in the last seven days
- Kimi K2 0905
Nous Portal$0.48 / $2 $0.60 / $2.5025% dearer · - Kimi K2 0905
Requesty$0.60 / $2.50 $0.57 / $2.307% cheaper · - Kimi K2 ThinkingNous Portal$0.48 / $2 $0.60 / $2.5025% dearer ·
Frequently asked
- Which is cheaper, Kimi K2 0905 or Kimi K2 Thinking?
- Kimi K2 0905, at $3 per million tokens against $3.41 for Kimi K2 Thinking, three parts input to one part output, read on 16 Sept 2026.
- Where is each one cheapest?
- Kimi K2 0905: $3 at NanoGPT. Kimi K2 Thinking: $3.41 at Vercel AI Gateway.
- Can I buy both from one provider?
- Yes. 17 providers sell both: Helicone, NanoGPT, OpenCode Zen, Vercel AI Gateway, TrustedRouter, io.net Intelligence, and more. The cheapest of them for the two together is Helicone.
- Which holds more context?
- Both hold 262k tokens in one call.
Prices exclude tax and are per million tokens, read 16 Sept 2026 from each provider's own catalog. The standard price blends three parts input to one part output; a page is 600 tokens in and 60 out. Bars are scaled to the larger value in each row.