OpenRouter(@OpenRouterAI)

Use it now: https://t.co/gtW9Ly7f4E Read more:

OpenRouter (@OpenRouter) — Mon, 19 Oct 2026 16:54:07 GMT

Use it now: openrouter.ai/inception/merc…

Read more:

Stefano Ermon@StefanoErmon

1/4 Today, we’re releasing Mercury 3. Mercury 3 is by far our most powerful model and the most intelligent diffusion LLM ever released. It outperforms GPT-6 Luna and Haiku 5.5 on quality while running at 1000+ tok/sec.

Your browser does not support the video tag.

🔗 View on Twitter

🔗 View Quoted Tweet

💬0🔄0❤️15👀4029📊1

⚡ Powered by xgo.ing

Mercury 3 is live on OpenRouter! @_inception_ai's new diffusion model generates 1,000+ tok/s with 1...

OpenRouter (@OpenRouter) — Mon, 19 Oct 2026 16:54:03 GMT

Mercury 3 is live on OpenRouter!

@_inception_ai's new diffusion model generates 1,000+ tok/s with 1M context, 4x more than Mercury 2.5. Inception reports it beats GPT-6 Luna and Haiku 5.5 on IFBench, GPQA Diamond, and SWE-Bench Atlas.

50% off at launch: $0.10/M in, $0.35/M out

💬28🔄25❤️579👀27663📊101

⚡ Powered by xgo.ing

OpenRouter (@OpenRouter) — Mon, 19 Oct 2026 12:06:00 GMT

Read more:

MiniMax (official)@MiniMax_AI

🔗 View Quoted Tweet

💬0🔄0❤️7👀3840📊1

⚡ Powered by xgo.ing

Use it now: https://t.co/4k5Eep837W

OpenRouter (@OpenRouter) — Mon, 19 Oct 2026 12:01:13 GMT

Use it now: openrouter.ai/minimax/minima…

💬2🔄1❤️15👀4629📊4

⚡ Powered by xgo.ing

Stealth model reveal 🐰 Space Bunny was OpenRouter’s #1 model for 10 straight days, processing 56.7...

OpenRouter (@OpenRouter) — Mon, 19 Oct 2026 12:01:04 GMT

Stealth model reveal 🐰

Space Bunny was OpenRouter’s #1 model for 10 straight days, processing 56.7T tokens! The most of any stealth model we’ve run.

Space Bunny hopped off OpenRouter on Oct 5. It's back today as @MiniMax_AI's M3.1 Flash, and it's free for the next 72 hours.

💬53🔄71❤️957👀107256📊168

⚡ Powered by xgo.ing

Classify: use the Decisions API to automatically detect an agent stuck in a loop, and halt or nudge ...

OpenRouter (@OpenRouter) — Sat, 17 Oct 2026 16:00:17 GMT

Classify: use the Decisions API to automatically detect an agent stuck in a loop, and halt or nudge it on thresholds you set.

openrouter.ai/docs/cookbook/…

💬7🔄0❤️12👀4485📊8

⚡ Powered by xgo.ing

Looking for a weekend project? Two new OpenRouter cookbooks on @scaledown_team's task-specific model...

OpenRouter (@OpenRouter) — Sat, 17 Oct 2026 16:00:14 GMT

Looking for a weekend project? Two new OpenRouter cookbooks on @scaledown_team's task-specific models (currently 20% off!), each a ready-to-run recipe for a workflow problem.

Summarize: compact a 100-step agent session while keeping the facts it still needs.

openrouter.ai/docs/cookbook/…

💬19🔄6❤️113👀13165📊35

⚡ Powered by xgo.ing

DeepSeek V4.1 Flash 1017 is live on OpenRouter with DeepSeek as the first provider and many more pro...

OpenRouter (@OpenRouter) — Sat, 17 Oct 2026 08:42:59 GMT

DeepSeek V4.1 Flash 1017 is live on OpenRouter with DeepSeek as the first provider and many more providers on the way.

@deepseek_ai's update to V4.1 Flash is more token efficient and shows major gains in agentic benchmarks.

Use it now: openrouter.ai/deepseek/deeps…

💬40🔄64❤️1063👀58983📊152

⚡ Powered by xgo.ing

Perplexity Decider v1.2 and Decider Flash v1.2 are now on OpenRouter @perplexity_ai's open-weights ...

OpenRouter (@OpenRouter) — Fri, 16 Oct 2026 20:58:00 GMT

Perplexity Decider v1.2 and Decider Flash v1.2 are now on OpenRouter

@perplexity_ai's open-weights decision models take text or images and return typed answers with probabilities

- Decider v1.2: a new 27B checkpoint, succeeding v1.1
- Decider Flash v1.2: 4B, faster and half the cost

openrouter.ai/perplexity/ppl…openrouter.ai/perplexity/ppl…

💬11🔄8❤️120👀45856📊25

⚡ Powered by xgo.ing

Images go in the same state as your text, up to 4 per request. See more: https://t.co/drqJzvdqHS

OpenRouter (@OpenRouter) — Fri, 16 Oct 2026 19:57:28 GMT

Images go in the same state as your text, up to 4 per request.

See more:
openrouter.ai/respan/span-02

💬0🔄0❤️6👀2295📊1

⚡ Powered by xgo.ing

It runs on the OpenRouter Decisions API, the same request format as the other decision models we hos...

OpenRouter (@OpenRouter) — Fri, 16 Oct 2026 19:57:25 GMT

It runs on the OpenRouter Decisions API, the same request format as the other decision models we host, so trying Span-02 is a one-line model change.

💬2🔄0❤️9👀3020📊3

⚡ Powered by xgo.ing

Span-02 from @RespanAI is now on OpenRouter An open-weight decision model: give it text and images,...

OpenRouter (@OpenRouter) — Fri, 16 Oct 2026 19:57:22 GMT

Span-02 from @RespanAI is now on OpenRouter

An open-weight decision model: give it text and images, define your questions, and get back a choice, a yes/no/unknown, or a score, with probabilities.

$0.01/M input tokens, free output, 131K context.

Respan@RespanAI

Span-02 is the cheapest decision model: $0.01/M input tokens.

With thinking off, it delivers frontier-level decision-making at lowest latency.

Need more? Give it a reasoning budget. Span-02 thinks until the deadline, then returns its best decision so far. Stop reasoning anytime.

Or let Span-02 decide. In auto mode, an adaptive learned gate picks the reasoning budget for every decision.

Fully open weights with a rust inference engine. 7× faster than vLLM.

🔗 View Quoted Tweet

💬12🔄9❤️64👀5922📊16

⚡ Powered by xgo.ing

Data live on our rankings page now: https://t.co/PYLvlPtX1P

OpenRouter (@OpenRouter) — Fri, 16 Oct 2026 18:25:23 GMT

Data live on our rankings page now:

openrouter.ai/rankings#open-…

💬1🔄1❤️4👀2567📊2

⚡ Powered by xgo.ing

Open-weight models processed 72% of tokens across OpenRouter last week. Open-weight models have see...

OpenRouter (@OpenRouter) — Fri, 16 Oct 2026 18:25:19 GMT

Open-weight models processed 72% of tokens across OpenRouter last week.

Open-weight models have seen more than 50% of tokens across the platform since mid-April, but the majority of customer spend remains on closed-weight models.

💬6🔄9❤️89👀7935📊17

⚡ Powered by xgo.ing

It runs on the same Images API as every other image model on OpenRouter, so moving your edits to Kre...

OpenRouter (@OpenRouter) — Fri, 16 Oct 2026 12:00:43 GMT

It runs on the same Images API as every other image model on OpenRouter, so moving your edits to Krea 3 is a model ID change.
See more: openrouter.ai/krea/k3-edit

💬1🔄0❤️1👀2347📊1

⚡ Powered by xgo.ing

Krea 3 leaves alone what you didn't ask it to touch: build a new scene around a product and the labe...

OpenRouter (@OpenRouter) — Fri, 16 Oct 2026 12:00:40 GMT

Krea 3 leaves alone what you didn't ask it to touch: build a new scene around a product and the label stays sharp, or give it character references and keep the same cast across a whole set.

💬1🔄0❤️5👀2665📊2

⚡ Powered by xgo.ing

Krea 3 is live on OpenRouter @krea_ai's new image editing model takes 1–4 reference images plus a p...

OpenRouter (@OpenRouter) — Fri, 16 Oct 2026 12:00:27 GMT

Krea 3 is live on OpenRouter

@krea_ai's new image editing model takes 1–4 reference images plus a prompt and returns an edited image at native 2K, most edits in under 8 seconds end to end
openrouter.ai/krea/k3-edit

Your browser does not support the video tag.

🔗 View on Twitter

💬15🔄2❤️70👀9373📊23

⚡ Powered by xgo.ing

Small vision-language models (9B total, 2B active MoE): image + text in, 32K context, optional reaso...

OpenRouter (@OpenRouter) — Thu, 15 Oct 2026 23:23:05 GMT

Small vision-language models (9B total, 2B active MoE): image + text in, 32K context, optional reasoning. Use /caption {"length":"normal"} for captions

Both are $0.30/M input and $2.50/M output tokens

See more:
openrouter.ai/moondream/moon…

💬1🔄0❤️8👀3458📊2

⚡ Powered by xgo.ing

Moondream 3 Preview and Moondream 3.1 from @moondreamai are now on OpenRouter, served by @PrimeIntel...

OpenRouter (@OpenRouter) — Thu, 15 Oct 2026 23:23:02 GMT

Moondream 3 Preview and Moondream 3.1 from @moondreamai are now on OpenRouter, served by @PrimeIntellect

Object detection and pointing work in plain chat: attach one image and start your message with /detect {"object":"coffee cup"} or /point {"object":"croissant"}

💬9🔄8❤️80👀20556📊21

⚡ Powered by xgo.ing

Run Claude Code (or your favorite harness like Codex, Pi, and more) on any model on OpenRouter: $ c...

OpenRouter (@OpenRouter) — Thu, 15 Oct 2026 22:24:44 GMT

Run Claude Code (or your favorite harness like Codex, Pi, and more) on any model on OpenRouter:

$ curl -fsSL openrouter.ai/labs/ori/insta… | bash
$ ori claude --model google/gemini-4-argon

Get started today at: openrouter.ai/ori/harness

💬2🔄0❤️4👀2789📊3

⚡ Powered by xgo.ing

Claude Opus 5.5, Gemini 4 Argon and GPT-6 Astra each rebuilt the same Royale: Last Agent Standing's ...

OpenRouter (@OpenRouter) — Thu, 15 Oct 2026 22:24:41 GMT

Claude Opus 5.5, Gemini 4 Argon and GPT-6 Astra each rebuilt the same Royale: Last Agent Standing's game visuals.

All three models ran in Ori Claude on OpenRouter through Ori Harness - giving you access to 500+ models across your favorite harnesses.

jacky@jjacky

the gpt-3 moment for game dev is here

a few months ago i built royale: last agent standing

royale drops 11 models into a battle royale to test long term agent planning+actions

models have gotten WAY better at game dev, so i asked 3 models to rebuild royale's visual engine

🧵

🔗 View Quoted Tweet

💬10🔄3❤️106👀12135📊15

⚡ Powered by xgo.ing

Any agent doing inference through OpenRouter can opt in to Armor by adding it as a plugin with custo...

OpenRouter (@OpenRouter) — Thu, 15 Oct 2026 21:00:37 GMT

Any agent doing inference through OpenRouter can opt in to Armor by adding it as a plugin with custom rules: [{ id: "armor" }].

Point your agent at the Armor docs to learn more: openrouter.ai/docs/guides/fe…

💬0🔄0❤️31👀4790📊7

⚡ Powered by xgo.ing

On an internal test set we generated for this task, Jev 1.13 had the highest accuracy and lowest cos...

OpenRouter (@OpenRouter) — Thu, 15 Oct 2026 21:00:34 GMT

On an internal test set we generated for this task, Jev 1.13 had the highest accuracy and lowest cost when compared to other Decisions models and Claude Code's usage of Sonnet 5.

💬4🔄2❤️39👀11561📊9

⚡ Powered by xgo.ing

Today, we are launching OpenRouter Armor in beta. Tool calls allow agents to interact with the worl...

OpenRouter (@OpenRouter) — Thu, 15 Oct 2026 21:00:20 GMT

Today, we are launching OpenRouter Armor in beta.

Tool calls allow agents to interact with the world, and regex-based filtering often isn't enough to detect unwanted tool calls.

Armor uses @typesafeai's Jev model to guard against out of policy tool calls before your agent executes them, with better performance than Claude Code's auto classifier at a fraction of the cost.

Your browser does not support the video tag.

🔗 View on Twitter

💬81🔄65❤️1157👀148284📊290

⚡ Powered by xgo.ing

New addition to our data rankings: Open vs Closed model share. Open-weight models processed 73% of ...

OpenRouter (@OpenRouter) — Thu, 15 Oct 2026 19:29:37 GMT

New addition to our data rankings: Open vs Closed model share.

Open-weight models processed 73% of total tokens so far this week.

Dance through the data yourself at openrouter.ai/rankings

We are also accepting suggestions for new public charts!

💬12🔄8❤️136👀14283📊27

⚡ Powered by xgo.ing

The upscaler takes a video you already have and now re-renders it in 4K. It's on OpenRouter for the ...

OpenRouter (@OpenRouter) — Thu, 15 Oct 2026 18:42:05 GMT

The upscaler takes a video you already have and now re-renders it in 4K. It's on OpenRouter for the next week in early access.

FLUX 3 Video can now make 4K videos, up from 1080p, from a prompt or an image for $0.41/s.

Try them out:
openrouter.ai/black-forest-l…openrouter.ai/black-forest-l…

Your browser does not support the video tag.

🔗 View on Twitter

💬2🔄1❤️5👀2177📊3

⚡ Powered by xgo.ing

FLUX 3 Image Klein is small enough to run on a single consumer graphics card, creates images in 4K, ...

OpenRouter (@OpenRouter) — Thu, 15 Oct 2026 18:41:54 GMT

FLUX 3 Image Klein is small enough to run on a single consumer graphics card, creates images in 4K, and starts at $0.009 per image.

openrouter.ai/black-forest-l…

💬1🔄0❤️6👀2628📊2

⚡ Powered by xgo.ing

New from @bfl_ai on OpenRouter today: - FLUX 3 Image Klein, a new open-weights image model that run...

OpenRouter (@OpenRouter) — Thu, 15 Oct 2026 18:41:48 GMT

New from @bfl_ai on OpenRouter today:

- FLUX 3 Image Klein, a new open-weights image model that runs on one consumer GPU
- An improved FLUX Video Upscale in early access, turns any video into 4K for $0.28/s (was $0.59/s)
- FLUX 3 Video can now create 4K videos from a prompt

💬8🔄3❤️59👀6916📊14

⚡ Powered by xgo.ing

Launching day 0 on OpenRouter with @baseten @FireworksAI_HQ @nebiusai and @togethercompute https://...

OpenRouter (@OpenRouter) — Thu, 15 Oct 2026 18:02:09 GMT

Launching day 0 on OpenRouter with @baseten@FireworksAI_HQ@nebiusai and @togethercomputeopenrouter.ai/meta/muse-spar…

💬0🔄0❤️5👀2350📊1

⚡ Powered by xgo.ing

Muse Spark Open from @AIatMeta is now available on OpenRouter! - strongest open-weight model from a...

OpenRouter (@OpenRouter) — Thu, 15 Oct 2026 18:02:06 GMT

Muse Spark Open from @AIatMeta is now available on OpenRouter!

- strongest open-weight model from a US lab
- built for agentic coding
- 1M context
- fully multimodal: image, video, audio and PDF input
- $1.25/M in, $4.25/M out

AI at Meta@AIatMeta

Today we’re releasing Muse Spark Open, the most capable US open-weight model to date.

Built for coding and agentic work, it excels at long-horizon tasks, complex software engineering, and multi-step reasoning.

Key Model Specs:
→ 816B total / 42B active parameters
→ 1M token context window
→ Natively multimodal across text, image, video, audio

🔗 View Quoted Tweet

💬16🔄20❤️258👀16605📊35

⚡ Powered by xgo.ing

So far in our private beta, we've seen teams running dedicated GLM, Kimi, and DeepSeek deployments, ...

OpenRouter (@OpenRouter) — Thu, 15 Oct 2026 16:45:18 GMT

So far in our private beta, we've seen teams running dedicated GLM, Kimi, and DeepSeek deployments, plus reserved OpenAI and Anthropic capacity.

Read more about how to try it: openrouter.ai/blog/announcem…

💬1🔄2❤️6👀2299📊2

⚡ Powered by xgo.ing

Private Endpoints are live on OpenRouter. Add dedicated capacity and fine-tunes that can be called ...

OpenRouter (@OpenRouter) — Thu, 15 Oct 2026 16:45:14 GMT

Private Endpoints are live on OpenRouter.

Add dedicated capacity and fine-tunes that can be called just like any other model. They're only visible to your org and automatically fall back to shared capacity when needed.

Available on Enterprise plans today.

Your browser does not support the video tag.

🔗 View on Twitter

💬12🔄6❤️93👀8394📊24

⚡ Powered by xgo.ing

Try it: https://t.co/nMwAvm5wm3

OpenRouter (@OpenRouter) — Thu, 15 Oct 2026 15:30:21 GMT

Try it: openrouter.ai/nvidia/nemotro…

💬0🔄0❤️8👀2507📊1

⚡ Powered by xgo.ing

Hybrid Mamba-Transformer MoE with multi-token prediction for faster output. Built for document intel...

OpenRouter (@OpenRouter) — Thu, 15 Oct 2026 15:30:18 GMT

Hybrid Mamba-Transformer MoE with multi-token prediction for faster output. Built for document intelligence, web research, customer service, and video understanding.

💬2🔄0❤️4👀2741📊2

⚡ Powered by xgo.ing

NVIDIA Nemotron 3.5 Super VL is live on OpenRouter. An open 120B-A12B multimodal model from @NVIDIA...

OpenRouter (@OpenRouter) — Thu, 15 Oct 2026 15:30:15 GMT

NVIDIA Nemotron 3.5 Super VL is live on OpenRouter.

An open 120B-A12B multimodal model from @NVIDIAAI for long-running agents that see, reason, and act across text, images, documents, GUIs, and video. 1M context.

💬13🔄6❤️219👀13699📊34

⚡ Powered by xgo.ing

The full (still boring) release notes from TypeSafe: https://t.co/wxTQJIBsKS

OpenRouter (@OpenRouter) — Thu, 15 Oct 2026 00:16:03 GMT

The full (still boring) release notes from TypeSafe:

typesafe.ai/blog/jev-1-14

💬0🔄0❤️12👀4028📊1

⚡ Powered by xgo.ing

Jev 1.14 is now on OpenRouter. Per @typesafeai this release is "beautifully boring": • Everything: ...

OpenRouter (@OpenRouter) — Thu, 15 Oct 2026 00:16:00 GMT

Jev 1.14 is now on OpenRouter.

Per @typesafeai this release is "beautifully boring":
• Everything: a tiny bit better
• Multilingual: the biggest upgrade
• Multimodal: it can see now (experimental)
• Price: unchanged ($0.042/M input, output free)
• Drama: none

openrouter.ai/typesafe/jev-1…

💬26🔄36❤️686👀40914📊104

⚡ Powered by xgo.ing

Universal-3.6 Pro from @AssemblyAI is now live on OpenRouter Speech-to-text through AssemblyAI's Sy...

OpenRouter (@OpenRouter) — Wed, 14 Oct 2026 22:06:00 GMT

Universal-3.6 Pro from @AssemblyAI is now live on OpenRouter

Speech-to-text through AssemblyAI's Sync API. Compared with Universal-3.5 Pro:

- 30% fewer errors on short conversational utterances
- 12% fewer errors on average across supported languages
- Better recognition and formatting of phone numbers, addresses, emails, and codes
- 14 new languages: Korean, Catalan, Galician, Russian, Romanian, Estonian, Persian, Afrikaans, Marathi, Cantonese, Zulu, Urdu, Xhosa, and Norwegian Nynorsk

Try it now at only $0.30/hr

openrouter.ai/assemblyai/uni…

💬5🔄5❤️56👀7588📊12

⚡ Powered by xgo.ing

How it works: put the parts in state the same way you would in Chat Completions (image_url, input_au...

OpenRouter (@OpenRouter) — Wed, 14 Oct 2026 21:00:26 GMT

How it works: put the parts in state the same way you would in Chat Completions (image_url, input_audio, video_url), ask a choice question, get probabilities back. The whole request above was 801 input tokens, $0.00012.

Read more at: openrouter.ai/cloudflare/cle…

💬0🔄1❤️8👀2645📊1

⚡ Powered by xgo.ing

We fed Clef Omni a photo, 3 seconds of sizzle, a 2-second clip of a hot dog stand, and the word "lun...

OpenRouter (@OpenRouter) — Wed, 14 Oct 2026 21:00:23 GMT

We fed Clef Omni a photo, 3 seconds of sizzle, a 2-second clip of a hot dog stand, and the word "lunch?"

It said: hot dog, 96.5%.

Audio + video input for Decisions is live on OpenRouter. Four modalities, one call, probabilities instead of tokens. $0.15/M input, output free.

Your browser does not support the video tag.

🔗 View on Twitter

Cloudflare@Cloudflare

We are expanding the Clef decision model family with Clef-omni, natively processing audio, video, images, and text in a single pipeline. We’ve also lowered Clef-flash pricing and boosted Clef inference speeds by up to 2.0x. cfl.re/4s6nYeD

🔗 View Quoted Tweet

💬8🔄8❤️119👀10817📊21

⚡ Powered by xgo.ing

ElevenLabs had a strong launch week, taking nearly a third of spend on text-to-speech models yesterd...

OpenRouter (@OpenRouter) — Wed, 14 Oct 2026 16:15:02 GMT

ElevenLabs had a strong launch week, taking nearly a third of spend on text-to-speech models yesterday after only 7 days live on platform.

All @ElevenLabs models are 50% off, only on OpenRouter, through 10/19

💬6🔄6❤️68👀6785📊13

⚡ Powered by xgo.ing

3/ Pick from 10 voices out of the box, or bring your own HeyGen key to use voices cloned in your own...

OpenRouter (@OpenRouter) — Wed, 14 Oct 2026 15:28:09 GMT

3/ Pick from 10 voices out of the box, or bring your own HeyGen key to use voices cloned in your own HeyGen account.

See more: openrouter.ai/heygen/heygen-…

💬0🔄0❤️4👀2145📊1

⚡ Powered by xgo.ing

2/ Pricing, 50% off through October 31: HeyGen Voice: $15 per million characters (list $30) HeyGen V...

OpenRouter (@OpenRouter) — Wed, 14 Oct 2026 15:28:06 GMT

2/ Pricing, 50% off through October 31:
HeyGen Voice: $15 per million characters (list $30)
HeyGen Voice Turbo: $7.50 per million characters (list $15)

💬1🔄0❤️11👀2548📊2

⚡ Powered by xgo.ing

1/ HeyGen Voice is now on OpenRouter, 50% off through October 31. @HeyGen's new text-to-speech mode...

OpenRouter (@OpenRouter) — Wed, 14 Oct 2026 15:28:03 GMT

1/ HeyGen Voice is now on OpenRouter, 50% off through October 31.

@HeyGen's new text-to-speech model topped the Artificial Analysis TTS leaderboard. It comes in two modes: HeyGen Voice, and the lower-latency HeyGen Voice Turbo.

HeyGen@HeyGen

HeyGen is now #1 in voice.

Our new voice model, HeyGen Voice, just topped the Artificial Analysis TTS leaderboard, beating every major model in blind tests.

Live on HeyGen + API.

API users get 50% off through 10/31. $15/million characters instead of $30, applied automatically.

Your browser does not support the video tag.

🔗 View on Twitter

🔗 View Quoted Tweet

💬13🔄13❤️110👀13221📊22

⚡ Powered by xgo.ing

OpenRouter users are loving Anthropic's Opus 5.5 It is the #1 model in share of spending across the...

OpenRouter (@OpenRouter) — Tue, 13 Oct 2026 19:00:41 GMT

OpenRouter users are loving Anthropic's Opus 5.5

It is the #1 model in share of spending across the platform and has led a rebound for Anthropic in the competition with OpenAI

Sonnet 5.5 and Haiku 5.5 have also seen strong launch periods

💬31🔄16❤️309👀17473📊61

⚡ Powered by xgo.ing

Try it now: https://t.co/cLnK9l7fy6

OpenRouter (@OpenRouter) — Tue, 13 Oct 2026 18:00:03 GMT

Try it now: openrouter.ai/liquid/lfm-2.5…

💬2🔄1❤️9👀2859📊3

⚡ Powered by xgo.ing

LFM2.5-24B-A2B from @liquidai is now on OpenRouter. A 24B MoE with 2B active parameters that reason...

OpenRouter (@OpenRouter) — Tue, 13 Oct 2026 18:00:00 GMT

LFM2.5-24B-A2B from @liquidai is now on OpenRouter.

A 24B MoE with 2B active parameters that reasons before it answers, built for agents and tool use.

Liquid AI@liquidai

Today we release 𝙻𝙵𝙼𝟸.𝟻-𝟸𝟺𝙱-𝙰𝟸𝙱, the largest 𝙻𝙵𝙼𝟸.𝟻 family member. It is an efficient local model for laptops and PCs to work through multi-step tasks inside popular agent harnesses.

> 24B total parameters, 2B active per token
> Context expanded to 131K tokens
> A 128K-token vocabulary
> 33T pretraining tokens
> It is a reasoning model
> Very easy to fine-tune and customize
> Performs competitively with Qwen3.6-35B-A3B, a model with ~50% more total and active parameters, on agentic and tool-use tasks.
> Open base and post-trained weights on @huggingface  🤗

1/

🔗 View Quoted Tweet

💬9🔄7❤️135👀10376📊18

⚡ Powered by xgo.ing

Mercury Decide from @_inception_ai now accepts image input on OpenRouter Put images alongside text ...

OpenRouter (@OpenRouter) — Tue, 13 Oct 2026 17:25:28 GMT

Mercury Decide from @_inception_ai now accepts image input on OpenRouter

Put images alongside text in the state of a Decisions request, and Mercury Decide answers your typed questions from what's in the image: a choice, a score, or a yes/no with a probability attached

Pricing is unchanged: free endpoint, or $0.02/M input with zero data retention. Images bill as input tokens

openrouter.ai/inception/merc…

💬6🔄11❤️82👀8649📊17

OpenRouter(@OpenRouterAI)

Use it now: https://t.co/gtW9Ly7f4E Read more:

Mercury 3 is live on OpenRouter! @_inception_ai's new diffusion model generates 1,000+ tok/s with 1...

Read more:

Use it now: https://t.co/4k5Eep837W

Stealth model reveal 🐰 Space Bunny was OpenRouter’s #1 model for 10 straight days, processing 56.7...

Classify: use the Decisions API to automatically detect an agent stuck in a loop, and halt or nudge ...

Looking for a weekend project? Two new OpenRouter cookbooks on @scaledown_team's task-specific model...

Read more from @deepseek_ai

DeepSeek V4.1 Flash 1017 is live on OpenRouter with DeepSeek as the first provider and many more pro...

Perplexity Decider v1.2 and Decider Flash v1.2 are now on OpenRouter @perplexity_ai's open-weights ...

Images go in the same state as your text, up to 4 per request. See more: https://t.co/drqJzvdqHS

It runs on the OpenRouter Decisions API, the same request format as the other decision models we hos...

Span-02 from @RespanAI is now on OpenRouter An open-weight decision model: give it text and images,...

Data live on our rankings page now: https://t.co/PYLvlPtX1P

Open-weight models processed 72% of tokens across OpenRouter last week. Open-weight models have see...

It runs on the same Images API as every other image model on OpenRouter, so moving your edits to Kre...

Krea 3 leaves alone what you didn't ask it to touch: build a new scene around a product and the labe...

Krea 3 is live on OpenRouter @krea_ai's new image editing model takes 1–4 reference images plus a p...

Small vision-language models (9B total, 2B active MoE): image + text in, 32K context, optional reaso...

Moondream 3 Preview and Moondream 3.1 from @moondreamai are now on OpenRouter, served by @PrimeIntel...

Run Claude Code (or your favorite harness like Codex, Pi, and more) on any model on OpenRouter: $ c...

Claude Opus 5.5, Gemini 4 Argon and GPT-6 Astra each rebuilt the same Royale: Last Agent Standing's ...

Any agent doing inference through OpenRouter can opt in to Armor by adding it as a plugin with custo...

On an internal test set we generated for this task, Jev 1.13 had the highest accuracy and lowest cos...

Today, we are launching OpenRouter Armor in beta. Tool calls allow agents to interact with the worl...

New addition to our data rankings: Open vs Closed model share. Open-weight models processed 73% of ...

The upscaler takes a video you already have and now re-renders it in 4K. It's on OpenRouter for the ...

FLUX 3 Image Klein is small enough to run on a single consumer graphics card, creates images in 4K, ...

New from @bfl_ai on OpenRouter today: - FLUX 3 Image Klein, a new open-weights image model that run...

Launching day 0 on OpenRouter with @baseten @FireworksAI_HQ @nebiusai and @togethercompute https://...

Muse Spark Open from @AIatMeta is now available on OpenRouter! - strongest open-weight model from a...

So far in our private beta, we've seen teams running dedicated GLM, Kimi, and DeepSeek deployments, ...

Private Endpoints are live on OpenRouter. Add dedicated capacity and fine-tunes that can be called ...

Try it: https://t.co/nMwAvm5wm3

Hybrid Mamba-Transformer MoE with multi-token prediction for faster output. Built for document intel...

NVIDIA Nemotron 3.5 Super VL is live on OpenRouter. An open 120B-A12B multimodal model from @NVIDIA...

The full (still boring) release notes from TypeSafe: https://t.co/wxTQJIBsKS

Jev 1.14 is now on OpenRouter. Per @typesafeai this release is "beautifully boring": • Everything: ...

Universal-3.6 Pro from @AssemblyAI is now live on OpenRouter Speech-to-text through AssemblyAI's Sy...

How it works: put the parts in state the same way you would in Chat Completions (image_url, input_au...

We fed Clef Omni a photo, 3 seconds of sizzle, a 2-second clip of a hot dog stand, and the word "lun...

ElevenLabs had a strong launch week, taking nearly a third of spend on text-to-speech models yesterd...

3/ Pick from 10 voices out of the box, or bring your own HeyGen key to use voices cloned in your own...

2/ Pricing, 50% off through October 31: HeyGen Voice: $15 per million characters (list $30) HeyGen V...

1/ HeyGen Voice is now on OpenRouter, 50% off through October 31. @HeyGen's new text-to-speech mode...

Read more insights from 600+ trillion monthly tokens in Inferred, our data newsletter. Subscribe for...

OpenRouter users are loving Anthropic's Opus 5.5 It is the #1 model in share of spending across the...

Try it now: https://t.co/cLnK9l7fy6

LFM2.5-24B-A2B from @liquidai is now on OpenRouter. A 24B MoE with 2B active parameters that reason...

Mercury Decide from @_inception_ai now accepts image input on OpenRouter Put images alongside text ...