MiniMax H3 Highspeed and H3 Lite from @MiniMax_AI are now available on Merge Gateway. H3 Lite costs half what H3 Max does at 768p ($0.04/sec) and $0.025/sec at 480p, with 5-second clips back in 8 to 9 seconds. H3 Highspeed is 25% under H3 Max at $0.06/sec. Check it out:

We're almost at capacity for next Wednesday's evening jazz social. @Tiny_Fish , @merge_api , @MongoDB , @vorfluxai , and Fuse Al are taking over Sohn, a popular Korean-inspired cafe near Dogpatch. DJ, unlimited drinks, and pastries. Founders, heads of partnerships, and engineering leaders from some of our favorite companies in the Bay are on the list. If you're a founder building Al infrastructure, GTM tools, or Al for recruiting, register below.

We spent yesterday at the Chief AI Officer Summit by AI Accelerator Institute (@AI_AInstitute), where @Shensi gave a keynote on designing AI tools that employees will actually use. Thanks to everyone who came to the talk and stopped by our booth!

Two models released in the past week now beat Claude Sonnet 5.5 (@AnthropicAI) at agentic coding. Kimi K3.1 Pro (@kimi_moonshot) passes 67.2% of Terminal-Bench 4.0, and Gemini 4 Carbon (@GoogleDeepMind), out yesterday, passes 65.2%. Kimi K3.1 Pro has the same token price as Sonnet ($2 in, $10 out per million) and costs about a third as much to run AA's Intelligence Index ($3.68 a task vs $10.02). Scores via week of Oct 26.

Particle (@particle_api) has been one of our closest partners in building Merge Gateway, serving a big share of our GLM 5.3 Flash and DeepSeek V4 Flash traffic. Today they launched particle-1, their first model for developers. It keeps performance at parity while using ~45% fewer tokens. Particle's inference and particle-1 are both available on Merge Gateway now. Try it out:

The Parallel Web Systems Decisions (@p0) API, a decision engine with world knowledge built in, is now available on Merge Gateway. Get responses grounded in real-time web data in under a second at $0.001 per request. Use it for any decision that depends on today's news, from flagging a risky vendor to picking an agent's next step. Try it out:

Across 420 runs per server, agents hit tool errors in 40 runs on Merge versus 70 on native, 43% fewer. We tested Merge against Microsoft's MCP servers across Outlook, OneDrive and SharePoint. Merge also exposed 128 tools compared to 96. 🧵🧵

At a two hour event, you only have time to really get to know five people. As a host, you're judged by the connections that last beyond the room. Every event we run is curated by industry, funding stage, and ICP. Every accepted guest naturally belongs and has a reason to come beyond the invite. @Tiny_Fish is hosting "Off the Clock" on November 4th at Sohn, a Korean-inspired cafe near Dogpatch with our co-hosts @merge_api , @vorfluxai , Fuse AI, and @MongoDB . We bought out the cafe for a jazz co-working night for founders and engineering leaders building in AI infrastructure, GTM, and recruiting. Bring your laptop and get some work done, or spend the evening getting to know others in the room. We’ll have unlimited drinks, pastries, and jazz music. It’s going to be a good time. Link to register below. Excited to host with our friends again!

Meet the eight newest connectors in Merge Agent Handler! - Meta Ads (@Meta) - AWS Partner Central Selling (@awscloud) - SAP S/4HANA Cloud (@SAP) - Zoom Phone (@Zoom) - TikTok for Business (@TikTok_US) - GoHighLevel (@gohighlevel) - Ironclad (@ironclad_inc) - Dialpad (@Dialpad) Try them out:

We're at booth 23 at @_odsc AI West in Burlingame through Wednesday. Come grab a hat and enter to win our Mac mini giveaway! Booking a demo gets you two entries for the Mac mini on our table, and we'll draw the winner at the booth Wednesday at 3:10pm. You MUST be present to win. Stop by for a demo and grab a "Context please" hat before they run out!

First official evals ran on our Context Layer V2 product coming soon... Vs. using MCP alone: - accuracy gains of ~40% - latency reduction by ~30% - cost savings of nearly 50% It's insane what a tiger team can accomplish

GPT-6.1 Terra from @OpenAI is now available on Merge Gateway. Matches GPT-6 Astra at one-tenth the price per token, and finishes tasks over 3× faster on average than Astra and GPT-6.1 Sol. Check it out:

This Thursday, Oct 29, we'll be at the Chief AI Officer Summit by @AI_AInstitute in Boston! @Shensi will be giving a keynote on product thinking for workforce AI, and how to design AI tools that employees will actually use. Event details:

Only four labs score above 50 on the Intelligence Index: @AnthropicAI (57.6), @Meta (55.8), @OpenAI (52.7) and @GoogleDeepMind (52.6). The next-best lab, SpaceXAI with Grok 4.7, scores 46.5, 6 points lower. MiMo-V2.6-Pro (@Xiaomi) is the top open-weights model, just behind at 46.3. Merge Gateway gives you every lab's best model through one API, so switching when the leader changes takes a config change. Scores via week of Oct 26.

Kimi K3.1 Pro from @Kimi_Moonshot is now available on Merge Gateway. It's Moonshot's most capable open-weight model yet, with a 1M-token context window and full-project coding across Unreal, Unity, and Godot. Check it out:

Next Tuesday and Wednesday we'll be at ODSC (@_odsc) AI West in San Francisco, booth #23#. Come by if you're connecting agents to the tools your customers already use. Event details: