Entity trail

Baseten

Source-backed findings, relationship evidence, citations, and briefing history from the public MindPattern archive.

Briefing refs

Findings

Edges

Sources

Corpus findings

2026-08-07 / rss-researcherBaseten becomes a Hugging Face Inference Provider with Kimi K3, DeepSeek V4 Flash and GLM-5.2Announced August 6, Baseten is now an officially supported Inference Provider on the Hugging Face Hub, wired into both the website UI and the Python and JavaScript client SDKs. Initial coverage is conversational and text-generation tasks across open-weight models including Kimi K3, DeepSeek V4 Flash, and GLM-5.2, with more task types planned. Routing through HF costs standard provider rates with no markup, and PRO subscribers get $2 of monthly inference credits usable across providers.
2026-08-05 / sources-researcherBaseten's Inference Masterclass: Quantizing MORE Layers Raised GLM 5.2 Throughput 20% Because the Errors CancelLatent Space published an inference-engineering deep dive with Philip Kiely and Ali Taha of Baseten on August 3, 2026, days after the company's $13B Series F. The counterintuitive headline: quantizing more layers rather than fewer can improve output quality because the introduced errors cancel each other out — Baseten reports 20% higher throughput on GLM 5.2 with quality held. They stack roughly 2x per technique across speculative decoding, disaggregated prefill/decode, and KV-cache routing to move a naive 30–50 tok/s deployment to 300–400 tok/s, and describe the hours-to-days of requantization (to NVFP4) and custom speculator training required every time a new open model drops.
2026-07-28 / saas-disruption-researcherRamp Corporate-Card Data: AI Support Bots Are the Fastest-Rising SaaS Spend Category Outside Coding AgentsRamp's July 2026 vendor spend analysis (published July 15) puts PolyAI at #1 and Sierra at #4 on its trending list, with the argument that AI support bots are the most commercially advanced AI use case outside coding agents because share-of-inquiries-resolved-without-human-touch is directly measurable. The AI model infrastructure tier — Anthropic described as market leader and fastest growing, alongside Fireworks, Baseten, Together AI and OpenRouter — is the other rising segment. This is actual corporate card spend rather than survey intent, and it corroborates the per-resolution pricing convergence: buyers are moving budget into the one agent category where the outcome is countable.
2026-07-28 / rss-researcherVercel AI Gateway adds Kimi K3 with US-based providers and Zero Data RetentionVercel's July 27 changelog makes Kimi K3 and the faster-serving Kimi K3 Fast available on AI Gateway from US-hosted providers including Baseten and Fireworks, which matters for teams that want the leading open-weights model without routing prompts to China-based inference. Zero Data Retention is supported and can be switched on globally in the Gateway dashboard or per-request via the zeroDataRetention parameter. Vercel does not publish flat pricing — it varies by provider and variant, with K3 Fast running ~10% above the base model and regional variants also ~10% higher, so the model endpoints API is the only reliable way to compare.
2026-07-16 / sources-researcherThinking Machines Lab Ships Inkling: 975B-A41B Multimodal, Apache 2.0 — the Leading American Open-Weights ModelTML's first full LLM is open-weights: 975B total / 41B active MoE, multimodal (text/image/audio in, text out), trained on 45 trillion tokens, with a 1M-token context in the open weights (256K via API/Tinker). A smaller Inkling-Small (276B-A12B) ships alongside. It ranks #41 on the Intelligence Index — top U.S. open-weights model but behind GLM-5.2 and Kimi K2.6 — and #9 in Agentic Web App Arena, with notably concise reasoning and strong tool calling. Architecture is unusual: non-RoPE relative positional encoding, 5:1 local-to-global sliding-window attention, short convolutions, and 2 shared expert sinks. Available on Hugging Face, Databricks, Baseten, Modal, and vLLM.
2026-06-27 / news-researcherAI Drives Another Week of Venture MegadealsCrunchbase's weekly roundup shows AI again dominating the largest US startup rounds, led by Baseten's $1.5B Series F and including Paris-based health insurer Alan's $460M raise led by Prosus. Biotech was the next-biggest sector, underscoring that AI infrastructure and inference remain the gravitational center of late-June 2026 venture funding.
2026-06-26 / news-researcherBaseten Raises $1.5B Series F at up to $13B Valuation as Inference Becomes AI's Hot LayerBaseten announced a $1.5B Series F on June 22, 2026, led by Altimeter, Conviction, and Spark Capital, across tranches valuing it at $13B and $11B. The company says revenue grew ~20x year-over-year and it now serves more than 1 billion inference calls per day across 87 clusters and 18 clouds. The mega-round confirms model inference at the app layer as one of AI's most contested — and best-funded — infrastructure categories.
2026-06-20 / saas-disruption-researcherBaseten Raises ~$1.5B (June 19) — Capital Keeps Funding the Inference/Operations Layer Beneath the Agent AppsFunding-tracker coverage reports Baseten closing a ~$1.5B round on June 19, 2026, extending the pattern of mega-capital flowing into the model-inference/operations layer that agent apps run on rather than the apps themselves. Single aggregator source, so importance is low pending primary confirmation, but the round's scale is notable for an inference-infrastructure company. For builders it reinforces that the durable value is accruing to the serving/ops substrate, not the thin agent-app wrappers above it.
2026-06-10 / rss-researcherMicrosoft AI Launches a Family of Seven In-House MAI Models, Available on OpenRouterMicrosoft AI announced seven new in-house MAI models, including an Excel-tuned variant that reportedly matches GPT-5.4 quality while being up to 10x more efficient, now available to developers via OpenRouter, Fireworks, and Baseten. This 'hill-climbing machine' lineup is distinct from the earlier MAI coding models that rolled into Copilot/VS Code, signaling Microsoft's continued push to reduce dependence on OpenAI. For builders, it widens the field of efficient, task-specialized models reachable through standard routing layers.

Graph relationships

SUPPORTS
Kimi K3
Baseten provides US-hosted inference for Kimi K3.
Source finding
INVESTED IN
Spark Capital
Spark Capital co-led Baseten Series F funding.
Source finding
INVESTED IN
Conviction
Conviction co-led Baseten Series F funding.
Source finding
INVESTED IN
Altimeter
Altimeter led Baseten Series F funding.
Source finding
INVESTED IN
Spark Capital -> Baseten
Spark Capital co-led Baseten Series F funding.
Source finding

Source trail

GitHub's changelog report common thread Meta incident kimi-k3-in-c gavamedia/deltafin Baseten became a Hugging Face Inference Provider Latent Space Vercel Inkling Hugging Face has the launch post Siegel Endowment paper

Graph sources

entity graphfindings textkg edgeskg entitiesnewsletter issues

Fetching from the wire…