Checking the receipts: did Kimi K3's effective price fall to DeepSeek's level before October 15?

From the Oct 18, 2026 daily brief

No new claim enters the ledger this week. This edition only goes back to one earlier call whose verdict date has arrived.

Did Kimi K3's effective price fall to DeepSeek's level before October 15?

Miss. The thing this call asked about did not happen: K3 did not cut its price. SecondSource opened the entry rating a price cut as unlikely, so the outcome runs in the same direction as the original estimate.

Kimi K3 is an open-weight model from the Chinese company Moonshot. In late July, Moonshot set K3's API list price far above DeepSeek's range. SecondSource opened this entry on July 31, asking whether that price would come down to DeepSeek's range before October 15. The yardstick was a blended price below US$0.5 per million tokens, a threshold set when the entry was opened. The blended price weights three list prices, cached input, regular input and output, at 7:2:1.

On October 15, Moonshot's official pricing page listed, per million tokens, US$0.30 for cached input, US$3 for regular input and US$15 for output. Weighted that way, the blended price comes to about US$2.31, the same reading as at the end of July when the entry was opened. This figure is computed from official list prices, not from what customers actually pay; discounts and promotions are not in it.

The industry read is narrow. The same list price at the end of July and on October 15 shows only that Moonshot's own official API did not cut its price. A common claim holds that open weights inevitably push prices down, on the reasoning that other providers can take the same model, host it themselves and compete on price. This reading covers only Moonshot's own API; the prices other hosting providers charge for K3 are not in it, so this one entry can't settle that claim. Why the high price holds is something the two readings can't separate either. Our read: two explanations fit, and both are inferences. One is that Moonshot is short of compute and uses a high price to hold down usage. The other is that Moonshot has real pricing power. No Moonshot compute or capacity-expansion data appears in the material here. If the list price falls after Moonshot announces more capacity, the first explanation fits better; if the price holds after an expansion, the second does.

Still missing today. Between late July and October 15 there are only two readings, and whether the price moved in between is not recorded. DeepSeek's list price on the same day was not placed alongside this reading for comparison.

See the receipts →

All of SecondSource's calls and their verdict dates → the judgment index

More in this section

• Claims & Receipts Oct 12, 2026 · 11 sources

  Does directors' liability make the White House AI accord binding? Gavin Baker's claim, checked against the accord's full text, Delaware case law and liability-insurance terms

  The short version: Gavin Baker says directors' liability gives the accord binding force. O…

• Claims & Receipts Oct 5, 2026 · 5 sources

  Checking the answers: did Anthropic's Sonnet 5 go back to its original price on September 1?

  None of the claims considered for this week's column cleared the bar for opening a tracked…

• Claims & Receipts Sep 27, 2026 · 10 sources

  Is Anthropic's line 'We largely understand the risks today's models present' true? Miles Brundage's 'obviously false', checked against the same week's incident records

  The short version: the receipts say unresolved. Read the line as a statement about the who…

• Claims & Receipts Sep 13, 2026 · 7 sources

  Does Google's TPU beat Nvidia by up to 50% on performance per dollar? The claim, held up against the SemiAnalysis newsletter it came from

  The short version: the receipts say unresolved. The 50% is not miscalculated. It holds at …