Pinned Tweet
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 ter performance than Claude Code's auto classifier at a fraction of the cost.
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 ter performance than Claude Code's auto classifier at a fraction of the cost.
81
68
1,171
153,788
Kimi K3.1 Pro from @Kimi_Moonshot is live on OpenRouter.
A 1M context, multimodal open-weight model, post-trained from Kimi K3 for long-horizon coding and agentic work. $2/M input tokens, $10/M output tokens.
Try it: openrouter.ai/moonshotai/kim…
A 1M context, multimodal open-weight model, post-trained from Kimi K3 for long-horizon coding and agentic work. $2/M input tokens, $10/M output tokens.
Try it: openrouter.ai/moonshotai/kim…
Introducing Kimi K3.1 Pro: our most capable 3T-class open-weight model.
🔹 1M-token context, 2.8T total / 104B active params (MoE), Multimodal.
🔹 Full-project coding: games in Unreal, Unity & Godot.
🔹 Excels at parametric 3D modeling in Blender & CAD.
🔹 More reliable at finance with hallucination down by 65%.
🔹 Matching the strongest proprietary models across coding, agentic and knowledge benchmarks.
Kimi K3.1 Pro and K3.1 Pro Highspeed (2–3× faster) are now live on Kimi.ai, Kimi Work, Kimi Code, and the Kimi API.
Open weights by Nov 4, 2026.
🔗 API: platform.kimi.ai
🔗 Tech blog: kimi.ai/blog/kimi-k3-1…
33
16
376
16,311
Moonshot reports 65.7% on Terminal-Bench 4.0, next to Claude Opus 5.5 (66.4%) and ahead of GPT-6 Astra (58.2%). On its Kimi Code Bench 3.0 Long-horizon, K3.1 Pro scores 61.2%, up from K3's 55.95%.
3
1
14
3,283
Kimi K3.1 Pro Highspeed is live too: 2-3x faster, at $4/M input and $20/M output. Pick it from the providers list on the model page.
See more:
openrouter.ai/moonshotai/kim…
See more:
openrouter.ai/moonshotai/kim…
0
0
11
2,973
Top models being used on security tasks, by spend %: GLM 5.3 & Kimi K3
Grok 4.7 is the top closed-weight model
openrouter.ai/rankings#task-…
Grok 4.7 is the top closed-weight model
openrouter.ai/rankings#task-…
16
10
260
15,302
OpenRouter retweeted
⚡ @OpenRouter tip #2
Use ~anthropic/claude-opus-latest as your model on OpenRouter and you are always on the newest Opus. When the next one ships, your code picks it up with no change.
There are aliases like this for most model families. What a few point to today:
-latest → Claude Opus 5.5
~openai/gpt-luna-latest → GPT-6 Luna
~google/gemini-flash-latest → Gemini 3.8 Flash
~moonshotai/kimi-latest → Kimi K3
Use ~anthropic/claude-opus-latest as your model on OpenRouter and you are always on the newest Opus. When the next one ships, your code picks it up with no change.
There are aliases like this for most model families. What a few point to today:
-latest → Claude Opus 5.5
~openai/gpt-luna-latest → GPT-6 Luna
~google/gemini-flash-latest → Gemini 3.8 Flash
~moonshotai/kimi-latest → Kimi K3
5
3
23
6,700
Use the Security Center to manage security across all models:
- turn on key spend alerts 🔔
- remove unused keys
- set a periodic spend limit on uncapped keys
- set a max key lifetime on ALL keys
… and more! openrouter.ai/settings/secur…
- turn on key spend alerts 🔔
- remove unused keys
- set a periodic spend limit on uncapped keys
- set a max key lifetime on ALL keys
… and more! openrouter.ai/settings/secur…
Today we're launching the first-ever gateway Security Center.
@OpenRouter now lets you audit all your keys in one place, see which ones are safe to remove, browse keys by level of risk, and control an IP allowlist.
9
4
58
8,577
Flint from @Recursive_SI is now on OpenRouter!
Recursive's first model: an open-weight 40B MoE (~2.5B active) built for coding agents by their autoresearch system, the Eureka Machine. 262K context, reasoning and tool calling.
Recursive reports 54.9% on Terminal-Bench 2.1, 89.7% on SWE-bench Pro v2 and 74.2% on SWE-bench Verified.
Free to use at launch.
Try it now: openrouter.ai/recursive/flin…
Recursive's first model: an open-weight 40B MoE (~2.5B active) built for coding agents by their autoresearch system, the Eureka Machine. 262K context, reasoning and tool calling.
Recursive reports 54.9% on Terminal-Bench 2.1, 89.7% on SWE-bench Pro v2 and 74.2% on SWE-bench Verified.
Free to use at launch.
Try it now: openrouter.ai/recursive/flin…
Introducing Flint, @Recursive_SI’s first model: a fast local coding model built with our automated research system, the Eureka Machine.
• Best-in-class fast local coding
• Up to 2x faster than comparable models: 125+ tokens/sec on a MacBook Pro, 150+ on RTX Spark
• Built from hs
Private, unmetered intelligence on your own machine. Open weights available now on @huggingface, run it locally on your favorite framework or try it for free on @opencode or @pidotdev.
17
19
232
19,765
Cache rates across the AI landscape have gotten markedly better in 2026.
We saw an 85% median cache rate in agentic requests across OpenRouter last month, up from 65% in the beginning of the year. And the vast majority of providers now serve agentic cache at 70% or higher.
We saw an 85% median cache rate in agentic requests across OpenRouter last month, up from 65% in the beginning of the year. And the vast majority of providers now serve agentic cache at 70% or higher.
23
6
111
12,949
Tip: see how different routers in the ecosystem benchmark using openrouter.ai/benchmarks/rou…
How different routers cost relative to frontier 👇
How different routers cost relative to frontier 👇
9
2
42
8,152
OpenRouter Auto costs about as much per token as our cheap-to-mid models, about 7x cheaper than Claude Opus 5.5 and 5x cheaper than Sonnet 5.5: openrouter.ai/openrouter/auto
3
0
8
2,881
Jev Router is about $0.65 per 1M, in the same range as GPT-6.1 Sol and Sonnet 5.5 and about 25% cheaper than Opus 5.5: openrouter.ai/typesafe/jev-r…
1
0
9
2,263
Agentic usage is up 34x since the start of February 2026.
Agentic usage is also now about 8x higher than human usage across OpenRouter on a weekly basis.
Agentic usage is also now about 8x higher than human usage across OpenRouter on a weekly basis.
15
18
157
19,304
Voxtral 3 from @MistralAI is now on OpenRouter!
Mistral released two new speech-to-text models and a new text-to-speech model, across 50+ languages.
Mistral released two new speech-to-text models and a new text-to-speech model, across 50+ languages.
Introducing Voxtral 3: state-of-the-art Transcribe and Speech models powering Mistral’s new voice agents API. 🧵
8
5
113
9,784
Voxtral Mini Transcribe 3: the smaller, lower-cost speech-to-text model in the family.
openrouter.ai/mistralai/voxt…
openrouter.ai/mistralai/voxt…
1
0
2
2,112
Voxtral TTS 3: voices that adjust emotion, pace and emphasis to the conversation, in 35+ languages and 50+ dialects.
openrouter.ai/mistralai/voxt…
openrouter.ai/mistralai/voxt…
0
0
3
2,051
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
@_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
48
33
743
46,018
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.
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.
61
79
1,109
131,049
Read more:
MiniMax M3.1 Flash: Space Performance + Bunny Speed
🚀🐰 Yes, Space Bunny - which has topped global leaderboards with 200T+ tokens - was an earlier version of M3.1 Flash.
💰Redefining the performance economics of agentic AI. M3.1 Flash achieves frontier-level intelligence s Cowork (Automation-Bench: 62%; ALE: 30%), Coding (DeepSWE: 74%) and Visual Reasoning (BabyVision: 91%), at 1/3 the cost of M3 and 3x the speed.
API pricing (flat rate): $0.15 / $0.60 per 1M input/output tokens, with cache reads at $0.015 / 1M tokens.
Model weights will arrive on Oct 23.
API: platform.minimax.io
M Plan: minimax.io/m-plan
MiniMax Code for desktop: agent.minimax.io/download
0
0
7
4,472