DeepInfra raises $107M Series B to scale the inference cloud — read the announcement

Updated November 2026.
Update, November 2026: Kimi K3.1 Pro is now available on DeepInfra. It’s post-trained from Kimi K3 with much larger post-training compute, swaps in sparse attention for long context, and costs $1.90 / $9.50 per 1M tokens on DeepInfra, a third below K3 (pricing as of 2026-11-17).
Moonshot AI’s Kimi K3 is the first open-source model to reach 2.8 trillion parameters, a scale that, until now, has been the exclusive territory of closed, proprietary systems. Built for long-horizon coding, agentic knowledge work, and multimodal reasoning, it activates 104 billion of those parameters per token through a sparse Mixture-of-Experts architecture, keeping inference tractable while sustaining frontier-level capability. On SWE-Marathon, a benchmark measuring sustained autonomous software engineering, K3 scores 42.0, ahead of every model in the comparison set, including Claude Fable 5 (35.0) and GPT-5.6 Sol (39.0).
What makes K3 architecturally interesting is how Moonshot got here: two attention mechanisms, Kimi Delta Attention (KDA) and Attention Residuals (AttnRes), combined with a Stable LatentMoE framework that selects 16 experts out of 896, deliver roughly 2.5x better scaling efficiency than its predecessor, Kimi K2. The model also ships with a 1-million-token context window and native vision support via MoonViT-V2, making it a credible option for the kind of multi-step, tool-using agent workflows where context depth actually matters. Full weights are available under the Kimi K3 License, meaning you can inspect, deploy, and build on top of the model, not just call an API.
And now, it’s available through DeepInfra.
Kimi K3 is the first open-source model to reach 2.8 trillion total parameters, described by both Moonshot and DeepInfra as the world’s first open 3T-class model. Despite that scale, only 104B parameters are active per token, achieved through a Mixture-of-Experts architecture that routes each token to 16 of 896 experts across a 93-layer network (1 dense layer, plus a mix of 69 Kimi Delta Attention layers and 24 Gated MLA layers). Training uses MXFP4 weights with MXFP8 activations via quantization-aware training, applied from the SFT stage onward.
Compared to Kimi K2 (the latest K2 release, Kimi K2.6, is also on DeepInfra), Moonshot reports roughly 2.5x improvement in overall scaling efficiency, the result of the KDA and Attention Residuals architecture combined with improvements in training methodology and data recipes.
Kimi K3 supports a 1,048,576-token context window natively, along with multimodal inputs (text, image, and video) via the MoonViT-V2 vision encoder (401M parameters). It’s built explicitly for long-horizon agentic tasks: extended coding sessions across large repositories, GPU kernel optimization, compiler development, vision-in-the-loop game development, CAD, and chip design, all coordinated through terminal tool use with minimal human oversight. For knowledge work, it produces deep research with interactive visualizations, widgets and dashboards, and motion design and video editing, powered by its native multimodal architecture.
On benchmarks, Kimi K3 sits just behind Claude Fable 5 and GPT-5.6 Sol overall, while outperforming Claude Opus 4.8 across most evaluated tasks. A few highlights:
| Benchmark | Kimi K3 | Claude Fable 5 | GPT-5.6 Sol | Claude Opus 4.8 |
|---|---|---|---|---|
| GPQA Diamond | 93.5 | 92.6 | 94.1 | 91.0 |
| SWE-Marathon | 42.0 | 35.0 | 39.0 | 40.0 |
| FrontierSWE | 81.2 | 86.6 | 71.3 | 66.7 |
| BrowseComp | 91.2 | 88.0 | 90.4 | 84.3 |
| MCPMark-Verified | 94.5 | 87.4 | 92.9 | 76.4 |
| OmniDocBench | 91.1 | 89.8 | 85.8 | 87.9 |
| Terminal-Bench 2.1 | 88.3 | 88.0 | 88.8 | 84.6 |
SWE-Marathon is worth calling out specifically: at 42.0, Kimi K3 leads every model in this comparison.
One thing to be aware of: Kimi K3 is a reasoning model, and Moonshot recommends passing the complete assistant message, including the reasoning trace, back on every turn in multi-turn conversations and tool calls. On DeepInfra you set reasoning depth with the reasoning_effort parameter (none, minimal, low, medium, high, xhigh or max), and “none” turns thinking off for simple requests. The model can also be proactive in agentic settings, taking autonomous decisions when facing ambiguity, so explicit constraints in the system prompt or an AGENTS.md file are worth setting up for bounded workflows.
Kimi K3 is available now on DeepInfra under the model identifier moonshotai/Kimi-K3 as a public endpoint, with a private endpoint deployment option also available. Pricing is straightforward: $2.85 per 1M input tokens, $14.25 per 1M output tokens, and $0.285 per 1M cached input tokens, a 10x discount on cache hits (pricing as of 2026-11-17). That’s 5% below Moonshot’s own API rate of $3.00 / $15.00, and DeepInfra’s output and cached-input prices are below those of every provider Artificial Analysis lists for Kimi K3. For long agent sessions, prompt cache retention keeps cached context for 5 minutes ($3.5625 per 1M tokens written) or 1 hour ($5.70 per 1M tokens written), and reuse within that window bills at the cached rate.
Output can use the full 1,048,576-token context window. If you leave max_tokens unset, DeepInfra uses the smaller of the remaining context and 65,536 tokens, so set it explicitly for long reasoning runs. The model supports JSON output, structured output, function calling, and multimodal inputs out of the box.
DeepInfra gives you access to Kimi K3 through an OpenAI-compatible API with zero setup and usage-based pricing, no infrastructure to provision, no cluster to manage. DeepInfra operates under a zero-retention policy and is SOC 2 and ISO 27001 certified.
Here’s everything you need to make your first call:
curl "https://api.deepinfra.com/v1/openai/chat/completions" \
-H "Content-Type: application/json" \
-H "Authorization: Bearer $DEEPINFRA_API_KEY" \
-d '{
"model": "moonshotai/Kimi-K3",
"messages": [
{
"role": "user",
"content": "Explain the tradeoffs between MoE and dense transformer architectures."
}
]
}'import os
from openai import OpenAI
client = OpenAI(
api_key=os.environ["DEEPINFRA_API_KEY"],
base_url="https://api.deepinfra.com/v1/openai",
)
response = client.chat.completions.create(
model="moonshotai/Kimi-K3",
messages=[
{
"role": "user",
"content": "Explain the tradeoffs between MoE and dense transformer architectures."
}
],
)
print(response.choices[0].message.content)import OpenAI from "openai";
const openai = new OpenAI({
apiKey: process.env.DEEPINFRA_API_KEY,
baseURL: "https://api.deepinfra.com/v1/openai",
});
const response = await openai.chat.completions.create({
model: "moonshotai/Kimi-K3",
messages: [
{
role: "user",
content: "Explain the tradeoffs between MoE and dense transformer architectures.",
},
],
});
console.log(response.choices[0].message.content);The only things that differ from a standard OpenAI call are the base URL (https://api.deepinfra.com/v1/openai), your DeepInfra API key (read from the DEEPINFRA_API_KEY environment variable), and the model name. The official OpenAI Python and Node.js SDKs work without modification.
Kimi K3 is a meaningful data point in the ongoing shift toward open models that can compete at the frontier, not just on paper benchmarks, but on the kind of sustained, multi-step work that real engineering workflows demand. The combination of a 1M-token context window, native vision support, and leading SWE-Marathon numbers makes it a serious candidate for teams building coding agents, research pipelines, or document-heavy applications where context depth and task continuity are non-negotiable.
Developers who get hands-on with K3 now will have a head start on understanding how large sparse models behave in production agentic settings, which is increasingly where the interesting problems live. Head to the Kimi K3 model page to run it in the playground or pull your API key and start building.
Qwen API Pricing Guide 2026: Max Performance on a Budget<p>If you have been following the AI leaderboards lately, you have likely noticed a name that keeps trading blows with the field’s proprietary leaders: Qwen. Developed by Alibaba Cloud, the Qwen model family has kept expanding at a rapid pace — the current lineup spans Qwen3, Qwen3.5, Qwen3.6, and the flagship Qwen3.8-Max, alongside dedicated coding […]</p>
OpenCode: Open-Source Claude Code Alternative<p>Open your cloud bill after a month of heavy agent use and the number stops being abstract. Teams report coding-assistant costs in the hundreds of dollars per developer, and some now set token budgets the way they once rationed cloud compute. Then in June 2026 the US government barred non-Americans from Anthropic’s Fable 5, and […]</p>
Wan 3.0: Model Documentation & Integration Guide<p>Wan3.0 is an all-in-one video generation model from Wan-AI, hosted on DeepInfra. It covers reference-based generation, editing, replication, and driving in a single model, and it accepts images, video clips, audio, documents, and web links as input. Use it for cinematic storytelling, social media content, and video editing. Every generation includes synchronized audio by default, […]</p>
© 2026 DeepInfra. All rights reserved.