Baseten vs fal

New in LangSmith: Engine v2, Managed Deep Agents, Fine-Tuning, and More

LangChain announced a wave of new LangSmith capabilities at Interrupt NYC, centered on LangSmith Engine v2, Managed Deep Agents v0.8, Trajectories, Fine-Tuning, and Custom Apps. Engine v2 adds red teaming, broader issue detection, and automated fix validation. Managed Deep Agents introduces identity-scoped auth and memory, expanded channels, and built-in web search. LangSmith Fine-Tuning turns agent trajectories into training data for open models, while Custom Apps lets teams build tailored interfaces on LangSmith data.

• LangChain announced the updates on September 25, 2026, at Interrupt NYC.
• LangSmith Engine v2 introduces red teaming, detection of performance trends and inefficient agent work, and automated validation of proposed fixes before human review.
• Managed Deep Agents v0.8 adds identity-scoped user memory, agent and user-owned credentials, Slack file transfer, HTTP channel support, and built-in web search powered by Parallel.

Goodfire Launches Internal AI Agent Monitors

Goodfire, a startup specializing in AI interpretability, launched on Thursday a new type of AI agent monitor that inspects a model's internal signals rather than reading its output, aiming to detect rogue behaviors more efficiently and at a fraction of the cost. The monitors are available to customers of Baseten, an AI model hosting platform. Baseten's Base Labs had previously announced a safety partnership with Goodfire and Hugging Face. Goodfire's approach uses small probes that scan a model's internal activations at each step, triggering a closer AI review only when flagged. In tests on the Kimi K3 model, Goodfire's monitors caught 94% of malicious hacking sessions and cost about $51 for 1,500 sessions, compared to $233 for a cheaper model and $10,000 for a top-tier one. The company positions the solution for open models, which can be stripped of safeguards, and sees it as critical for inference-time guardrails.

• Goodfire launched internal activation monitors for AI agents on October 8, 2026.
• Monitors are available to Baseten customers.
• In tests on Kimi K3, monitors caught 94% of malicious hacking sessions.

Announcing our partnership with OpenAI

Baseten announces a partnership with OpenAI, as featured on their blog.

Dubbing API: Video that Travels

VEED has launched its new Dubbing API, designed to translate video speech, clone the speaker's voice, and synchronize lip movements in the target language. Available in Standard and Pro tiers, the API supports up to 100 languages for videos up to 10 minutes long without requiring prior voice training. The API is accessible via VEED's self-serve platform starting October 28, 2026, with an integration on fal arriving in early November.

• Dubbing API translates speech, clones voices, and performs lip sync via a single API request for videos up to 10 minutes in length.
• Standard tier costs $0.11 per second of video and supports 100 languages.
• Pro tier costs $0.18 per second of video, supports 94 languages plus 14 accent variants, and handles multi-speaker videos while preserving background audio.

AI Inference Startup Fal Doubles Valuation to $16B

Fal, a San Francisco-based AI inference startup, has raised $1.3 billion in a Series F round led by Bessemer Venture Partners, with participation from Nvidia, Sound Ventures, HOF Capital, and T. Rowe Price. The company also announced a previously closed $350 million Series E round led by Sequoia Capital in April, bringing the combined total to $1.65 billion. The latest funding values Fal at $16 billion, doubling its April valuation of $8 billion. Founded in 2021 by Burkay Gur and Gorkem Yurtseven, Fal develops infrastructure for generative media, enabling businesses to run image, video, and audio AI models at scale across entertainment, advertising, e-commerce, gaming, and marketing. The funding will support expansion of its inference infrastructure and team, as well as new AI content creation approaches planned for 2027. Fal has surpassed $800 million in annualized revenue.

• Fal raised $1.3 billion in a Series F round led by Bessemer Venture Partners, with participation from Nvidia, Sound Ventures, HOF Capital, and T. Rowe Price.
• Fal announced a previously closed $350 million Series E led by Sequoia Capital, bringing combined Series E and F funding to $1.65 billion.
• The funding values Fal at $16 billion, double its April valuation of $8 billion.