Observed Signal · Jul 17, 2026 · Product Launch · Source: DEV Community · Impact: 3/5 · Sentiment: Neutral
Large Language Models (LLM) & AI Market: Kimi K3: $15 Output Price Not Whole Cost
Moonshot AI launched Kimi K3 (announced July 16, 2026), a 2.8‑trillion‑parameter Mixture‑of‑Experts model with a one‑million‑token context window, multimodal vision, and an OpenAI‑compatible API. Official API pricing is $0.30 per million cache‑hit input tokens, $3 per million uncached input tokens, and $15 per million output tokens. The model’s weights were promised (expected July 27, 2026) but were not downloadable at launch, and Moonshot has not disclosed the active parameter count per token (MoE routes reportedly activate 16 of 896 experts). Independent measurements report higher-than-average output verbosity (≈130M output tokens vs a 63M median in one evaluation), meaning task cost depends heavily on output volume and caching. The article recommends guarded testing, output caps, and cost‑per‑task evaluation before migrating production traffic.
Frontier LLM launch with large context window, multimodal features, and explicit API pricing; impacts cost and architecture decisions for AI product teams but is not a platform-level policy change.
Key Takeaways & Evidence Grounding
- Kimi K3 was launched (announced) July 16, 2026 by Moonshot AI.
- Moonshot AI states Kimi K3 is a 2.8 trillion parameter Mixture‑of‑Experts model with a one‑million‑token context window and native vision.
- Kimi K3 weights were not downloadable at launch; Moonshot said weights would be released by July 27, 2026 (commitment, not completed release).
- Official API pricing: $0.30 per 1M cache‑hit input tokens, $3 per 1M uncached input tokens, and $15 per 1M output tokens.
- Independent measurement (Artificial Analysis) reported ~130M output tokens in an evaluation versus a 63M median for comparable models, increasing cost per task.
Connected Companies & Entities
5 Entities mappedAnthropic
Foundation model company selling AI assistants and model APIs.
“If you want to compare Kimi with OpenAI, Anthropic, and Google through one OpenAI-compatible endpoint, that's roughly what TokenMix does....”
Moonshot AI
No-code AI platform for continuous ecommerce conversion optimisation.
“Moonshot AI's official Kimi K3 announcement confirms a 2.8-trillion-parameter Mixture-of-Experts model, a one-million-token context window, ...”
OpenAI
Foundation model company selling AI software, APIs and subscriptions.
“If you want to compare Kimi with OpenAI, Anthropic, and Google through one OpenAI-compatible endpoint, that's roughly what TokenMix does....”
Artificial Analysis
Independent AI model benchmarking and selection platform.
“Artificial Analysis currently gives K3 an Intelligence Index of 57, reports about 62 output tokens per second, and measures a 1.99-second ti...”
Search, video, adtech and cloud giant within Alphabet.
“If you want to compare Kimi with OpenAI, Anthropic, and Google through one OpenAI-compatible endpoint, that's roughly what TokenMix does....”
Ontology Mapping & Concepts
Track Real-Time Market Signals & Shifts
Set up custom watchlists to receive automated, evidence-grounded executive digests whenever material signals or shifts occur across your tracked landscape.