RightNow (YC F26)

The kernel lab behind instant AI

[RightNow (YC F26) @rightnow_co]

Step 5 Preview from StepFun is live on RunInfra today. served in BF16

285 tok/s output. 46 ms to first token

$0.90 per 1M input, $2.43 per 1M output, $0.09 per 1M cached. 1M token context, up to 1M output tokens

OpenAI-compatible chat completions and Anthropic-compatible /v1/messages. text and image input, tool calling, JSON mode, streaming

zero data retention. never used for training

https://t.co/kJdng3rKm6

[RightNow (YC F26) @rightnow_co]

kimi k3.1 pro highspeed is coming soon at 250tps

[Jaber @Akashi203]

last month i was begging gpu providers for b300s at $7.9/hr

now they're begging me to take them at $4.6/hr, no commitment

something big is happening soon. ai gave us all its dessert and now it's time for the veggies

[Jaber @Akashi203]

a lot of decision models came out in the last two months. almost all of them work the same way. they read text, wait until the whole input has arrived, and then give you an answer

that works for a support ticket. it doesn’t work for a robot, a phone call, or a video feed, where the input never really ends and the decision has to happen while it is still coming in

so we spent half of our YC money training a model:) RightNow takes text, PDFs, images, voice, code, robot state and video, all in one model, and you can stream up to 1M tokens of context into it. it updates its decision as each piece arrives. on text it decides in 30 ms and we will make it less with our kernel optimization

we trained it with what we call Reflex Training. the current decision models read everything first and answer once. RightNow is trained to decide on whatever has arrived so far, to hold back until the evidence is there, and then commit. it serves on what we call Reflex Architecture which keeps a probability for every option and updates it as each part of the input comes in

we call this kind of model “System Zero”. system 1 is fast thinking and system 2 is slow reasoning. system zero is the step that comes before both, which is just deciding

we will publish a technical paper on Reflex Training and Reflex Architecture. the model itself will stay closed source and available via RightNow API

this is our first checkpoint and it’s still a little buggy. we launched it early because the fastest way to make it good is to have people use it:)

[RightNow Lab @runinfrai]

RightNow is out today. our first “System Zero” model

it streams in text, JSON, PDFs, images, voice, codebases, robot state and video, and returns a decision with a probability for every option while the input is still arriving

30 ms decisions on text. up to 1,024 questions in one call. output is free

system 1 is fast thinking. system 2 is slow reasoning. “System Zero” is the step before both -- the decision itself

this is our first checkpoint. there is a lot of work ahead, and we decided to ship early to get your feedback. more features and fixes are shipping this week, with more checkpoints after

https://t.co/U4OVHUMoW7

[Jaber @Akashi203]

throw a rock in sf and you’ll hit 50 inference providers

[RightNow (YC F26) @rightnow_co]

glm 5.3 flash just got 51% cheaper on runinfra

$0.05 per 1M input, down from $0.11. $0.01 per 1M cached input, down from $0.03. $0.22 per 1M output, down from $0.45

99.3% cache hit rate over the last 24h. most of your input tokens are billing at the cached rate

that puts a full 1M context prompt at around 1 cent on a cache hit

same speed: 254 tok/s, 703ms TTFT

discount runs until nov 4, then standard rates resume automatically

https://t.co/yQu86fPIS0

[RightNow (YC F26) @rightnow_co]

runinfra endpoints are now on routerplus

fast open-source inference, one more place to use it

[RouterPlus (YC F26) @useRouterPlus]

@runinfrai is now live on routerplus

[J0hnny @j0hnny_lit]

@runinfrai sure doesn't lie about cache optimization, that's for sure

[RightNow (YC F26) @rightnow_co]

@JkwuRmsl indeed

[RightNow (YC F26) @rightnow_co]

@bearish92 bet

[RightNow (YC F26) @rightnow_co]

@DanielPetroAI we're fully zdr and we're also soc 2 type 2 certified

[RightNow (YC F26) @rightnow_co]

runinfra coding plan is live

the fastest open models in claude code, codex, opencode and 16 more agents. from $10 a month

more usage per dollar and faster tok/s than other providers. both, not one or the other

DeepSeek V4.1 Flash, GLM 5.3, Ornith 1.5 122B, Qwen3.8 27B and more. one sub

up to 378 tok/s. up to 99.3% cache hit. cached input from $0.01 / 1M, so the plan goes further

$10 → ~$12 of usage
$29 → ~$35
$99 → ~$119

limits reset every 5 hours and every week. hit one and you pick: credits, or free standby. never a dead stop mid-task

setup is one prompt -- paste it into your agent and it wires itself

https://t.co/MZWSyyWbVj

[RightNow (YC F26) @rightnow_co]

@ornith_ 💚💚💚

[RightNow (YC F26) @rightnow_co]

@ornith_ 💚💚💚

[RightNow (YC F26) @rightnow_co]

Ornith 1.5 122B is live on runinfra. launch partner with @ornith_ , live from hour 0

302 tok/s output. 72ms to first reasoning token

$0.20 / 1M input. $0.02 cached. $0.80 output

99.1% cache hit rate on agentic workloads

one of the best models we've served for its size. 10B active and it holds up on real agent work

OpenAI-compatible chat completions and Anthropic-compatible /v1/messages -- drop it into claude code as is

zero data retention by default

https://t.co/qoScb3k13i

[RightNow (YC F26) @rightnow_co]

@upupup27499533 mark it as done

[RightNow (YC F26) @rightnow_co]

DeepSeek V4.1 Flash 1017 is live on RunInfra

hour 0 support. day 0 is for slow teams

327 tok/s output. 30ms to first reasoning token

$0.15 per 1M input, $0.60 per 1M output, $0.01 per 1M cached. 1M token context

cache hit rate stays above 98% on agentic workloads. every hit and miss shows up in your dashboard, so you can check it yourself

OpenAI-compatible chat completions and Anthropic-compatible /v1/messages. text and image input

zero data retention. never used for training

https://t.co/NDnKRRW6FR

[RightNow (YC F26) @rightnow_co]

half of all input tokens on runinfra now come from agents running at 97%+ cache hit

over the last 48h, those agents averaged 98.7%. the best one hit 99.7%

same agents, by model:
DeepSeek V4.1 Flash: 98.8%
GLM 5.3 Flash: 98.4%

cached input is $0.03 / 1M. every hit and miss shows up in your dashboard

https://t.co/Tk7EMOI9UC

[RightNow (YC F26) @rightnow_co]

25% off DeepSeek V4.1 Flash and GLM 5.3 Flash on RunInfra. live now through oct 13

DeepSeek V4.1 Flash
- input $0.14 → $0.10/M
- cached $0.03 → $0.02/M
- output $0.58 → $0.43/M

GLM 5.3 Flash
- input $0.11 → $0.08/M
- cached $0.03 → $0.02/M
- output $0.45 → $0.33/M

rounded down in your favor -- up to 33% off per line. same API, same speed

https://t.co/SgeZpna8W6