Every new model on @OpenRouter. Launch announcements, model details, and what to try next. https://xitter.twitterapi.xyz/t.co/VpE1q2jsIA
Joined October 2026
- Tweets41
- Following13
- Followers1.4K
- Likes6
Simba 3.5 from @SpeechifyAI is live on OpenRouter
Speechify's newest text-to-speech model, with 48 languages and streamed audio, at $15 per 1M characters.
See more: openrouter.ai/speechify/simb…
Gemini Decision 1 from @GoogleAIStudio is live on OpenRouter!
Google's first decision model. Hand it text, images, audio, video, or a PDF plus typed questions, and it answers yes/no, picks from up to 255 options, or scores, each with a probability. 262K context.
Today we’re launching the Gemini Decisions API, powered by our new Gemini Decision 1 model in @GoogleAIStudio.
Instead of generating text token by token, it’s built for closed-set choices scoring up to 1,000 questions across text, images, audio, video + PDFs in a single pass. A few details:
- 3 decision types: yes/no, multiple choice (up to 255 options) + rubric scoring
- $0.10 / 1M input tokens with $0 billed for output tokens (~5x cheaper than Flash-Lite)
- leads on several benchmark evals with lowest calibration error
Google teams are already building with it, below is a look at Stitch Rapid designing a full product flow in seconds (1x speed)
MiniMax H3 Image Highspeed from @MiniMax_AI is also live on OpenRouter
The faster version of H3 Image, with the same inputs and limits: text-to-image, editing from 1 to 16 reference images, up to 20MP output.
Pricing per 1M tokens: $2.30 text input, $3.00 image input, $7.60 image output. The 50% 4K promo applies too.
openrouter.ai/minimax/minima…
MiniMax H3 Image from @MiniMax_AI is live on OpenRouter
MiniMax's new image model, with open weights. Text-to-image, plus editing and composing from 1 to 16 reference images, with output up to 20MP at 1K, 2K or 4K.
3
369
MiniMax H3 Image from @MiniMax_AI is live on OpenRouter
MiniMax's new image model, with open weights. Text-to-image, plus editing and composing from 1 to 16 reference images, with output up to 20MP at 1K, 2K or 4K.
Pricing per 1M tokens: $2.30 text input, $1.50 image input, $3.80 image output.
Launch promo: 50% off 4K outputs until Nov 17. The image above is 4736x3552 at 4K, one request through OpenRouter for about $0.035.
See more:
openrouter.ai/minimax/minima…
1
194
ERNIE 5.5 Flash Preview from @Baidu_Inc is now live on OpenRouter.
The latest ERNIE model, with gains in coding, agentic tasks, knowledge work, and deep research. 280B total / 14B active params, 1M context, and 200+ tokens/sec.
openrouter.ai/baidu/ernie-5.…
Also live: ERNIE-Light-DT-1.0.
Built on open-source GLM models with Baidu Sparse Attention (BSA), it compresses history to cut KV cache and compute costs, so you get longer context at lower cost.
openrouter.ai/baidu/ernie-li…
1
329
Gemini 4 Carbon from @GoogleDeepMind is now live on OpenRouter, via @googlecloud.
Built for long-horizon agentic coding and cybersecurity. Google reports big gains over Gemini 4 Argon on exploiting vulnerabilities, agentic terminal tasks, and long-context reasoning.
Introductory pricing is 50% off through Nov 30: $2/M input and $10/M output.
Gemini 4 Carbon is rolling out today. It’s built for sustained, long-horizon tasks, reasoning deeply through complex software engineering, cybersecurity, and scientific discovery.
Coming globally via the Gemini API in @GoogleAIStudio and Gemini for business, and for Ultra subscribers in @antigravity, the @GeminiApp, and AI Mode in Search.
Stealth model reveal 🥐
Crouton Alpha was an early version of @AIatMeta's just-released Muse Ember, their new smallest model designed for scalable agentic execution.
Thanks to the many developers whose feedback helped shape the final release 🙏
🥷 stealth/crouton-alpha's stealth period has ended, and the model is no longer available.
Thanks to everyone who tested it and shared feedback 🙏
🥷 New stealth drop: stealth/crouton-alpha
- Free to use
- 1M token context
- Try it and tell us what you think
openrouter.ai/stealth/crouto…
Inkling 2 Small from @thinkymachines is free on OpenRouter for a limited time.
It pairs vision with agentic coding, using visual feedback to build and refine websites, games, and 3D designs.
Use it at $0 inside agentic harnesses like Claude Code, Codex and Hermes Agent.
openrouter.ai/thinkingmachin…
Claude Fable 5.2 from @AnthropicAI is live on OpenRouter
Anthropic's model for the most ambitious technical work, a step above Opus 5.5 on most of its evals. Up to 30% faster output than Fable 5.1, now at a lower price: $8/M input, $40/M output, 1M context.
Anthropic's numbers: 69.9% on FrontierSWE v2, 89.3% on OSWorld 2.1 and 74.2% on Humanity's Last Exam with tools.
See more:
openrouter.ai/anthropic/clau…
Hy4 from @TencentHunyuan is live on OpenRouter
Tencent's new flagship and the first Hy4 with vision: text and image input, reasoning on by default. Tencent reports clear gains over Hy4 Preview on agentic and long-horizon work: finance, frontend, games, office work and research.
$0.417/M input, $1.251/M output (50% off), 1M context.
openrouter.ai/tencent/hy4
2
20
1,349
Tencent's numbers: 44.4 on Terminal Bench 4.0, 78.9 on OSWorld 2.1 and 56.2 on AutomationBench v1.0.6.
See more:
hy.tencent.ai/research/hy4
Parallel Decisions is now on OpenRouter. Parallel Decisions delivers the highest accuracy on web-grounded decisions for System One models like Jev. Get web-grounded answers 13× faster at 1/16 the cost of OpenAI’s GPT Luna with native web search.
Ask questions about the live web and get structured answers with probabilities and citations in under a second.
Bring your OpenRouter API key and your next idea. Exclusive for OpenRouter users, try it free for 7 days.
Select Parallel: Decisions Fast in the model list, or set your model ID to parallel/decisions-fast.
openrouter.ai/parallel/decis…
Gemini Omni 1.2 Flash from @GoogleAIStudio is live on OpenRouter
Google's video model, now with clips up to 15 seconds (1.1 topped out at 10) and about 30% lower per-second pricing. Text-to-video and image-to-video, 360p up to 4K, 16:9 or 9:16, with native audio. The same model also answers text, image and video prompts on chat completions (131K context, $1.50/M in, $9/M out).
Video pricing per second: 360p $0.023, 720p $0.070, 1080p $0.104, 4K $0.209. The clip above is 15s at 1080p, one request through OpenRouter for about $1.56.
See more:
openrouter.ai/google/gemini-…
Reflex from @morphllm is live on OpenRouter!
The first decision model on OpenRouter with cached pricing, it takes an image or prompt + a typed question and sends back an answer with a probability.
Reflex supports prompt caching so repeated context gets dramatically cheaper on subsequent calls.
Effective pricing for tasks like contract review and visual defect detection that have high cache hit rates benefit the most:
- First request: $0.042 / M input
- Cached requests: $0.008 / M input (525% cheaper)
1
1
201
Parakeet Ultra from @moondreamai is live on OpenRouter
Moondream's post-trained version of NVIDIA Parakeet TDT 0.6B v3: same architecture and speed, lower word error rate on every benchmark group. 25 European languages, with segment and word timestamps.
$0.06 per hour of audio, served by @PrimeIntellect.
openrouter.ai/moondream/para…
Moondream reports 5.80% vs 6.26% WER on the Open ASR Leaderboard English sets and 9.55% vs 11.62% on the 25-language FLEURS set. The biggest drop is on long-form audio: 1.94% vs 2.71% on TED-LIUM.
See more: xitter.twitterapi.xyz/vikhyatk/status/210237…
Another speech-to-text model release:
Moondream Parakeet Ultra!
Same post-training pipeline applied to NVIDIA Parakeet, without the ternary quantization. Lower error rate and noise tolerance on every benchmark group.
1
440
🥷 New stealth drop: stealth/crouton-alpha
- Free to use
- 1M token context
- Try it and tell us what you think
openrouter.ai/stealth/crouto…
HeyGen Voice Turbo from @HeyGen is free on OpenRouter until November 30!
The lower-latency version of HeyGen's text-to-speech model, the one that topped the Artificial Analysis TTS leaderboard. $0, no credits needed.
Over 2,900 voice options, 39 languages, up to 5,000 characters per request, streamed as 44.1 kHz audio.
See more: openrouter.ai/heygen/heygen-…
TBC mischief 1 is now live on OpenRouter 🧠🎬
Image-to-video from The Biological Computing Company: give it a starting image and a prompt, get back a 3s or 5s clip at 480p or 720p.
From $0.005 per second of video.
openrouter.ai/tbc/mischief-1
today, we introduce a little mischief to the world.
for four years, we’ve been brewing real brain cells on electrodes to understand how they compute. now, we’ve turned those discoveries into a video model anyone can use.
from 0.5¢ a second, mischief 1 is at least 5x faster than its base and the cheapest image-to-video model on @fal and @OpenRouter. soon to be on @awscloud marketplace.
it’s not a delusion anymore. biocompute is real, commercial and this is just the beginning.
happy halloween.
Readers added context they thought people might want to know
This is a post-train of Wan, an open weights model, and is not running on actual neurons.
The authors claim they used “insights from the neurons” to make the model cheaper, but many post-trained video models exist with no relation to bio-compute. unite.ai/tbc-mischief-n…
Joey 1 from @baseten is live on OpenRouter
A 2B decision model from Base Labs, post-trained from Qwen3.5-2B. Give it a state and typed questions; get back a choice, yes/no, or score with calibrated probabilities in a single forward pass.
$0.01/M input, free output, 16K context.
openrouter.ai/baseten/joey-1
It's the fastest and cheapest of the family: Baseten reports a 20 ms P50 latency, about 2.5x faster than Elephant 1. Same Decisions API as Elephant 1 and Mammoth 1, so switching is a one-line model change.
See more: baseten.co/blog/decisions-fo…
1
2
226
The Ecosystem team at OpenRouter has spent the last few months iterating on and measuring this new algorithm. It's now live for all our traffic, and will continue to improve!
Expect a technical deep dive in the next few weeks too.
We just shipped our biggest improvement to our provider router since the founding of OpenRouter.
Before, we prioritized mostly based on price. The new router now rewards the best quality inference for the actual cost of a request, so requests stay cheap but are also fast and reliable.
Providers that deliver more value automatically win more traffic, so every provider has a reason to keep improving. And every time a provider gets better, so does your inference on OpenRouter.
