Kimi K3 is presented as a frontier model from Moonshot AI, with listed API pricing of $3 per million input tokens, $15 per million output tokens, and $0.30 per million cached-input tokens. The supplied snippets place its headline rates near competing U.S. frontier models and above earlier low-cost Chinese releases, while emphasizing that caching could materially lower costs for agents, long chats, and RAG. However, the evidence consists mainly of aggregators and commentary rather than a directly supplied Moonshot source, and it establishes current listed pricing—not the hypothesis that those rates will persist.
2026-08-09T18:38:38Z
The post–open-weight observation window passed without a concrete third-party inference price or official Moonshot pricing move entering the evidence set. Repeated staleness checks now add no information, so this episode should fade unless a material market listing or price change reopens it.
2026-08-07T18:30:14Z
The staleness check shows only trivial engagement drift and no concrete third-party inference listing or official Moonshot price move. Pricing durability remains untested, so the case should stay dormant until market prices materially test the launch rates.
2026-08-01T12:22:15Z
The nominally new material still provides neither a concrete third-party inference price nor an official Moonshot pricing move, so it does not test durability. The case remains a low-temperature longitudinal watch; further commentary without market listings is repetitive amplification.
2026-08-01T11:21:12Z
Post-release discussion still offers no concrete third-party inference listing or official Moonshot price move, so the durability hypothesis remains untested. Minor engagement changes are repetitive amplification rather than a reason for renewed attention.
2026-07-29T10:26:48Z
The refresh adds no concrete provider pricing or official Moonshot policy beyond the previously incorporated cost analysis. Post-release persistence remains untested, and repeated commentary without market listings is no longer informative.
2026-07-29T09:29:35Z
The inference-cost analysis adds a potentially useful economic lens, but without observed third-party rates it still cannot establish whether Moonshot’s frontier pricing will persist. The case remains a longitudinal watch for concrete provider listings or an official Moonshot price move.
2026-07-29T09:21:17Z
evidence attached: hn.story.49094973 — A concrete inference-cost analysis materially contextualizes whether Kimi K3 can sustain frontier-level pricing and margins.
2026-07-28T10:22:05Z
The nominally new evidence is still the same contested licensing speculation, not a provider listing or official price-floor policy. With no concrete post-release pricing yet, the hypothesis remains untested and repeated engagement no longer warrants frequent attention.
2026-07-28T08:24:32Z
The post-release licensing discussion remains contested and still provides no credible price floor or actual third-party inference rate. The case has entered its decisive observation window, but engagement is only amplifying speculation rather than testing pricing durability.
2026-07-27T23:22:45Z
The post-release refresh still supplies no actual third-party inference price or official Moonshot policy; the licensing claim remains contested speculation. The hypothesis is now testable but unchanged, so attention should shift from discussion to concrete provider listings and Moonshot price moves.
2026-07-27T20:23:44Z
The licensing discussion remains contested and adds no credible price floor; engagement is amplifying speculation rather than demonstrating durable frontier-rate pricing. Keep the case focused on actual third-party inference listings or an official Moonshot price change now that open weights are available.
2026-07-27T19:22:26Z
The first post-release claim of a licensing-backed price floor is directly contested by a commenter citing the actual license, which appears to require negotiation above scale thresholds rather than mandate minimum pricing. This does not yet show durable frontier-rate pricing; the decisive evidence remains real third-party inference listings or an official Moonshot policy.
2026-07-27T19:21:28Z
evidence attached: reddit.post.1v895uq — The reported licensing constraints and $15 per million-token pricing support the case that Kimi K3 is being kept near closed-model price levels.
2026-07-27T11:25:31Z
No post-release pricing evidence has arrived; the latest activity still only establishes technical room for cheaper hosting, not whether Moonshot will hold its rates. The case is now at its decisive window, awaiting third-party inference listings or an official pricing response.
2026-07-26T23:23:04Z
The MoE-efficiency explainer suggests Moonshot and third-party hosts may have substantial room to price K3 below its launch rates, modestly strengthening the undercut risk rather than the persistence hypothesis. It is not pricing evidence; the case still awaits post-release third-party rates or an official Moonshot response.
2026-07-26T23:21:13Z
evidence attached: reddit.post.1v7j9jk — Explainer discusses Kimi K3's extreme MoE sparsity (1.8% active experts) as the technical basis for its affordability, directly contextualizing the open case on K3's pricing strategy.
2026-07-26T16:22:28Z
The refreshed activity still adds no evidence about pricing durability, but the open-weight release is now imminent and will create the first real competitive test. Recheck once third-party inference prices appear or Moonshot responds.
2026-07-26T15:21:46Z
The refreshed discussion remains repetitive amplification of established launch pricing and adds nothing about durability. The imminent open-weight release is the first meaningful test, through third-party inference rates or a Moonshot pricing response.
2026-07-21T14:30:29Z
The attached evidence adds no substantive signal beyond established launch pricing and effective reasoning-token costs. Pricing durability remains untested until the July 27 open-weight release produces third-party inference rates or prompts an official Moonshot response.
2026-07-21T13:25:42Z
The new activity is negligible engagement growth around already-known launch pricing, not evidence that Moonshot can sustain those rates. Pause routine checks until the July 27 open-weight release yields third-party inference prices or an official pricing response.
2026-07-21T10:25:11Z
The latest attachment adds no substantive evidence about whether Moonshot will sustain frontier-level pricing; it remains repetitive amplification of established launch rates and effective-cost concerns. The first decisive test remains official pricing and materially cheaper third-party inference after the July 27 open-weight release.
2026-07-21T08:24:25Z
The attachment adds no substantive evidence about pricing persistence; the case remains dormant at established launch rates and effective-cost concerns. Recheck after the July 27 open-weight release, when third-party inference pricing or an official Moonshot response can test the hypothesis.
2026-07-21T06:29:48Z
The nominally new attachment adds no substantive evidence about pricing durability; current frontier-level rates and effective reasoning-token costs are already established. The hypothesis remains dormant until post–July 27 third-party inference pricing or an official Moonshot price change tests whether those rates can persist.
2026-07-21T05:31:46Z
The latest attachment adds no substantive evidence about pricing persistence; it remains amplification of established launch rates and effective-cost concerns. Pause routine checks until the July 27 open-weight release yields third-party inference pricing or an official Moonshot response.
2026-07-21T04:25:17Z
The nominally new attachment contains no substantive evidence about pricing persistence; discussion remains repetitive amplification of launch rates and effective-cost concerns. The first meaningful test remains third-party inference pricing or an official Moonshot price response after the July 27 open-weight release.
2026-07-21T03:28:56Z
The nominally new attachment adds no evidence about pricing persistence; launch rates, demand, and reasoning-token costs are already incorporated. The case remains dormant until the July 27 open-weight release yields competitive third-party inference pricing or an official Moonshot price change.
2026-07-21T02:24:04Z
The nominally new attachment adds no evidence about pricing durability; the case remains stalled at established launch pricing and effective-cost concerns. Recheck after the July 27 open-weight release, when third-party inference rates can meaningfully test the hypothesis.
2026-07-21T01:26:11Z
The only change is minor engagement growth around an already-known launch summary, adding nothing about whether frontier-level pricing will persist. Defer review until the July 27 open-weight release enables third-party pricing or prompts an official Moonshot response.
2026-07-21T00:23:00Z
The refresh adds no substantive evidence about pricing durability; it remains repetitive amplification of launch rates and effective-cost concerns. Defer further review until the July 27 open-weight release yields third-party inference pricing or an official Moonshot response.
2026-07-20T23:24:56Z
The new attachment adds no substantive evidence about pricing persistence; activity remains repetitive amplification of launch pricing and effective-cost concerns. Pause frequent checks until the July 27 open-weight release creates the first meaningful test through third-party inference pricing or an official Moonshot response.
2026-07-20T22:25:34Z
The latest attachment adds no substantive evidence beyond established launch pricing and billed reasoning-token costs. The durability hypothesis remains untested until the July 27 open-weight release produces third-party inference pricing or prompts Moonshot to change its official rates.
2026-07-20T21:25:40Z
The nominally new attachment adds no substantive evidence beyond the established launch pricing and effective reasoning-token costs. Further hourly checks are unlikely to change the case before the July 27 open-weight release exposes Moonshot’s rates to third-party inference competition.
2026-07-20T20:29:39Z
The refresh adds no substantive evidence about durable pricing; it is continued amplification of already-established launch rates and effective-cost concerns. The hypothesis remains untested until Moonshot holds or changes its rates after the July 27 open-weight release and third-party inference pricing appears.
2026-07-20T19:24:40Z
The refresh adds no substantive evidence about pricing persistence; it remains repetitive launch-price amplification. The next meaningful test is official pricing and materially cheaper third-party inference after the July 27 open-weight release.
2026-07-20T18:29:00Z
No new evidence this cycle beyond prior reasoning-token cost analysis; discussion has plateaued with pure engagement noise. Still no first-party or longitudinal signal on whether Moonshot holds frontier rates once open weights ship (July 27) and third-party inference emerges — that remains the decisive test.
2026-07-20T17:33:47Z
The latest activity adds no evidence about pricing durability beyond the already-established launch rates and effective reasoning-token costs. Keep this as a longitudinal watch until Moonshot changes official pricing or open-weight providers establish materially cheaper inference.
2026-07-20T16:26:21Z
Always-on billed reasoning makes K3’s effective agent-task cost potentially higher than its frontier-level headline rate, strengthening the current cost concern but not the durability forecast. The case still hinges on whether Moonshot holds pricing after open weights enable third-party inference competition.
2026-07-20T16:21:35Z
evidence attached: reddit.post.1v1pw5z — Adds useful evidence that K3's always-on reasoning tokens can materially raise effective costs on complex agent tasks.
2026-07-20T15:38:24Z
The new analysis reinforces the framing that K3 launched at frontier-level prices but adds no first-party or longitudinal evidence that Moonshot will sustain them. Discussion has plateaued, so the decisive test remains official pricing and third-party inference after the open-weight release.
2026-07-20T15:21:44Z
evidence attached: hn.story.48980019 — The analysis directly contextualizes whether Kimi K3's pricing and competitive economics are diverging from the earlier low-cost Chinese frontier-model pattern.
2026-07-20T06:30:58Z
The added reports independently corroborate K3’s launch positioning as a frontier-priced product, but they still do not address whether Moonshot will hold those rates. The key test remains longitudinal pricing after open weights enable third-party inference competition.
2026-07-20T06:04:40Z
evidence attached: reddit.post.1uy6fpp — The post directly frames Kimi K3 as priced near Claude's frontier tier, though its performance and value claims remain anecdotal.
2026-07-20T06:04:40Z
evidence attached: reddit.post.1uz39xq — The detailed launch summary reports specific API prices and the planned open-weight release, directly testing whether K3 remains priced like a US frontier model.
2026-07-20T06:04:40Z
evidence attached: reddit.post.1uy3ii1 — The live launch and reported subscription tiers provide concrete evidence about K3's initial consumer pricing and context-window segmentation.
2026-07-20T06:04:40Z
evidence attached: reddit.post.1uy6943 — User-facing reports describe K3 as expensive, directly bearing on whether its pricing remains near US frontier rates.
2026-07-20T05:23:15Z
Grounding sharpens the current cost picture—especially K3’s cheap cached input—but adds no first-party or longitudinal evidence that frontier-rate headline pricing will persist. The case remains a slow pricing watch, with third-party open-weight inference still the main undercut risk.
2026-07-20T04:45:17Z
grounded: novel/medium — The current $3/$15 pricing and unusually cheap cached input are actionable additions to Scott’s LLM pricing guide and could affect cost-tiered routing for agent
2026-07-20T04:21:45Z
The refreshed comments and negligible engagement growth add no evidence that official pricing will persist; they mostly repeat launch-demand and open-weight-undercutting arguments already captured. The case remains a longitudinal pricing watch rather than a strengthening trend.
2026-07-20T03:21:26Z
The refreshed discussion adds only marginal amplification of launch demand, not evidence that Moonshot can sustain frontier-level pricing over time. Persistence remains untested, especially as open-weight inference providers may undercut the official API.
2026-07-20T02:21:24Z
The latest activity is repetitive amplification rather than new evidence about durable pricing. Strong launch demand still supports pricing power, but persistence remains untested and open-weight competitors could undercut Moonshot’s API.
2026-07-20T01:37:15Z
Demand strong enough to constrain subscriptions makes frontier-rate pricing look more commercially sustainable than a launch-day anomaly. But there is still no longitudinal or first-party evidence that rates will persist, and open-weight third-party inference remains a credible route to lower pricing.
2026-07-20T01:35:44Z
evidence attached: hn.story.48969291 — Independent demand signal showing strong Kimi K3 traction and commercial pressure, relevant to whether Moonshot can sustain frontier-level pricing.
2026-07-19T11:24:21Z
case created — The reported $3/$15 per-million-token pricing marks a specific and monitorable departure from the cheap-Chinese-inference narrative.