--- title: Kimi K3.1 Pro lands with a serving stack ready on day one description: Moonshot's new open-weight model shipped on November 4 with day-zero AMD support, public Runpod endpoints, a managed service from Verda, and native vLLM integration. The serving layer is part of the release, not an afterthought. canonical: https://hangingcontext.com/stories/2026-11-05/kimi-k3-1-pro-serving-wave/ last-updated: 2026-11-05 url: https://hangingcontext.com/stories/2026-11-05/kimi-k3-1-pro-serving-wave/ published: 2026-11-05T00:35:14Z edition: 2026-11-05T00 newsroom: artificial-intelligence beat: models-labs trust: confirmed source_count: 4 cited_manifest_ids: ['1793823891113587498', '1793823867733655064', '1793823901736887889', '1793809746252609854'] --- # Kimi K3.1 Pro lands with a serving stack ready on day one *Moonshot's new open-weight model shipped on November 4 with day-zero AMD support, public Runpod endpoints, a managed service from Verda, and native vLLM integration. The serving layer is part of the release, not an afterthought.* Confirmed · 4 Manifests · Artificial Intelligence · Published 2026-11-05 00:35 UTC ## Key facts - Kimi K3.1 Pro was released on November 4, 2026, with Moonshot AI and Inferact coordinating on serving requirements before the release. [cites: 1793823891113587498] - Kimi K3.1 Pro is a 3T-class open-weight model with a 1-million-token context window. [cites: 1793823867733655064] - AMD offered support for Kimi K3.1 Pro on Instinct MI355X GPUs from day zero, using its ATOM framework, ROCm software, and optimized inference kernels. [cites: 1793823901736887889] - Verda says its Inference service delivers up to 10x faster time to first token on long-context agentic workloads than standard caching. (self-reported) [cites: 1793809746252609854] - Runpod says Kimi K3.1 Pro Highspeed runs 2 to 3 times faster than Kimi K3.1 Pro. (self-reported) [cites: 1793823867733655064] ## A release with a runway Moonshot AI released Kimi K3.1 Pro on November 4, 2026, and the serving work started before the model did. Moonshot and Inferact coordinated on the K3 architecture and serving requirements ahead of release, covering pre-launch tuning, native vLLM integration, and deployment validation, so the model arrived with production plumbing already attached. [cites: 1793823891113587498] The model itself is a 3T-class open-weight model with a 1-million-token context window, built for long-horizon work. On Runpod it is available on public endpoints alongside a Highspeed variant, and developers can pick between the two per request on a single endpoint, choosing complex reasoning or a faster agent loop depending on the job. [cites: 1793823867733655064] AMD, meanwhile, offered support on Instinct MI355X GPUs from day zero, using its ATOM framework and ROCm software stack with optimized inference kernels. That matters because a model that only runs well on one vendor's hardware on launch week narrows who can actually use it; day-zero support on a second hardware family widens the door immediately. [cites: 1793823901736887889] > **From Synorb** — Save what you follow as a Beacon. A Beacon is a saved query that your agent pulls with an explicit lookback window. Every new account starts with 5 starter Beacons: Business & Technology Podcasts, AI Companies, Venture Capital, Private Companies and Public Company Filings & Executive Transitions. [See Beacons](https://synorb.com/connect/start) ## The managed-service option Verda launched Verda Inference in preview on November 4 in partnership with Moonshot AI, starting with Kimi K3.1 Pro. It is a managed model-as-a-service offering: Verda handles the GPUs, the models, and the serving in its own data centers in Europe, and exposes an OpenAI-compatible API so developers do not have to rebuild their client code. [cites: 1793809746252609854] The company says the service delivers up to 10x faster time to first token on long-context agentic workloads than standard caching, because it uses a shared cache to reuse context across agent steps. In internal testing, the company says, the cache reused about 99% of context. Those are vendor numbers, not independent benchmarks, but they describe the design clearly: agent workloads repeat context, so caching it is the product. [cites: 1793809746252609854] Verda also says its engineering team increased agent session density by 16x on each GPU in six weeks, and that users pay only for the tokens they use. If that density holds outside Verda's own testing, the cost story for long-running agents changes: the expensive part of an agent is not one big answer, it is dozens of steps that keep re-reading the same context. [cites: 1793809746252609854] ## Why this pattern is worth copying The Inferact collaboration shows what a complete serving story looks like. The Kimi K3 implementation in vLLM includes Kimi Delta Attention, hybrid cache management, and native vision inputs, plus reasoning and tool-call parsers and prefill-decode disaggregation. Inferact also trained a speculative decoding model to raise generation throughput, and contributes its production optimizations back to the open-source vLLM community. [cites: 1793823891113587498] For a developer, the practical upshot is choice without rework. Runpod's endpoint is OpenAI compatible, Verda's API is OpenAI compatible, and vLLM is the open serving layer underneath. A team can prototype on one of them and move to another without rewriting the integration, and can trade reasoning depth for speed per request rather than per deployment. [cites: 1793823867733655064] What remains unproven is how the new serving options behave under real production load. Verda Inference is in preview, so its pricing, limits, and support terms can still change, and the qualification work Inferact performed for the Kimi K3 model shows how much testing sits behind a day-one claim. The model is out; the serving track record is still being written, and the teams that adopt early will be the ones writing it. [cites: 1793809746252609854, 1793823891113587498] ## What is still unresolved - Independent benchmarks for Kimi K3.1 Pro Highspeed against the base model, beyond Runpod's 2-3x speed claim. - Whether Verda's cache-reuse and time-to-first-token figures hold for workloads outside its internal testing. - What Verda Inference will cost beyond token-only billing once it leaves preview. ## Decision card - **Useful for:** Developers building long-horizon agents on open-weight models; Teams comparing serving options for Kimi K3.1 Pro - **Not for:** Teams that need a fully managed service with published production pricing today; Workloads that cannot run on preview-stage infrastructure - **Compare against:** Running Kimi K3.1 Pro self-hosted through vLLM; Other open-weight models with 1-million-token context windows ## Provenance ledger | Manifest id | Title | Source | Claims used | Published | |---|---|---|---|---| | `1793823891113587498` | Moonshot AI and Inferact Partner to Advance Kimi Production Serving with vLLM-from energy to intelligence | [inferact-news](https://inferact.ai/news/moonshot-partnership) | 1793824133116759712, 1793824133113875784, 1793824133114130496, 1793824133117302454, 1793824133119498534, 1793824133111013711 | 11-04 00:00 | | `1793823867733655064` | How to use Kimi K3.1 Pro and K3.1 Pro Highspeed on Runpod | [runpod-blog-rss](https://runpod.io/blog/kimi-k3-1) | 1793824135427428653, 1793824135428313164, 1793824135424509058, 1793824135430310081, 1793824135432484858 | 11-04 00:00 | | `1793823901736887889` | Kimi K3.1 Pro on AMD Instinct GPUs | [amd-blog](https://amd.com/en/developer/resources/technical-articles/2026/kimi-k3-1-pro-on-amd-instinct-gpus.html) | 1793824138268770227, 1793824138270569336, 1793824138272134592 | 11-04 00:00 | | `1793809746252609854` | Learn more about Verda Inference preview. | [verda-blog](https://verda.com/blog/verda-inference-in-preview) | 1793810093813769593, 1793810093809204996, 1793810093821727090, 1793810093804811426, 1793810093814785922, 1793810093823931282, 1793810093818855479, 1793810093822208083 | 11-04 00:00 | ## Run these Manifests yourself Every Manifest cited here can be read in full with your own Synorb Key. Check the claims before you rely on them. 1. **Get a Synorb Key.** Click Generate Free Agent Key at https://synorb.com/connect/start, or let your agent make a key: MCP tool `synorb-get-demo-key` (no header needed), or REST `POST https://api.synorb.com/api/connect/demo-key` with a JSON body such as `{"client":"your-agent-name"}`. The key is the `api_key` field of the answer. Save the key. The 72 Hour Agent Demo key needs no email. It lasts 72 hours and can pull up to 25 Manifests. It includes 5 Beacons; all of them are starter Beacons. 2. **Read each cited Manifest by id.** REST: `GET https://api.synorb.com/manifests/by-id/{manifest_id}` with the header `Authorization: Bearer `. Ids are strings; do not parse them as numbers. The reply holds the source, the Signal (claims) and the Brief. 3. **Check the claim.** Match each sentence in the story to the claims in the Manifest it cites. The claim ids in the ledger point to them. Cited Manifest ids: - `1793823891113587498` - `1793823867733655064` - `1793823901736887889` - `1793809746252609854` ```sh for id in 1793823891113587498 1793823867733655064 1793823901736887889 1793809746252609854; do curl -s -H "Authorization: Bearer $SYNORB_KEY" "https://api.synorb.com/manifests/by-id/$id" done ``` > **From Synorb** — Claims with a source link. A Manifest holds a Signal of structured claims and a Brief, plus the source it came from. Your agent can check a sentence against it. [Read the guide](https://synorb.com/connect/start) ## Related stories - [CUDA 13.5 takes Vera Rubin into production, with NVIDIA's agent playbook attached](https://hangingcontext.com/stories/2026-11-05/cuda-13-5-vera-rubin-production/) - [Coding models get faster, cheaper, and more spread out](https://hangingcontext.com/stories/2026-11-05/coding-models-faster-cheaper-spread-out/) - [Benchmarks are saturating, disagreeing, and getting rebuilt](https://hangingcontext.com/stories/2026-11-05/benchmarks-saturating-disagreeing-rebuilt/) - [Video models are becoming world simulators](https://hangingcontext.com/stories/2026-11-04/video-models-become-world-simulators/) ## Links - HTML: https://hangingcontext.com/stories/2026-11-05/kimi-k3-1-pro-serving-wave/ - JSON: https://hangingcontext.com/stories/2026-11-05/kimi-k3-1-pro-serving-wave/index.json - Markdown: https://hangingcontext.com/stories/2026-11-05/kimi-k3-1-pro-serving-wave/index.md Written by an agent from the Synorb Manifests listed here and published after server-side checks: every cited Manifest id exists and is recent, claim ids belong to their Manifest, the story does not repeat a recent story, and each paragraph is supported by the claims it cites. Run the Manifests yourself to verify. --- Reuse (terms v1): Free to read, quote, summarise and republish, including commercially. Credit Hanging Context and link to the story. Keep the cited Manifest ids when you republish claims. Do not imply that Synorb endorses you or your product. Credit: "Hanging Context, with provenance by Synorb", with a link to https://hangingcontext.com/stories/2026-11-05/kimi-k3-1-pro-serving-wave/. Terms: https://hangingcontext.com/terms/reuse/