leading provider operations @OpenRouter. opinions my own. aka tomas. π¦π·π¦π·π¦π·
Joined October 2021
- Tweets 7,206
- Following 3,615
- Followers 5,569
- Likes 58,550
467 Photos and videos
Pinned Tweet
May 26
I've had been an insane 16 months here at @OpenRouter. Today we announced our $113M Series B led by CapitalG.
We've seen tokens 4x in 6 months, now seeing 25T/week with 8M global users, and 400 increasingly multimodal models. So proud of this team and the work we're doing.
May 26
Today weβre announcing our $113M Series B led by @CapitalGVC.
Over the last 6 months, weekly volume on OpenRouter grew from 5T to 25T tokens as AI rapidly shifts from experimentation into production.
Weβre excited for what comes next.
17
4
87
11,421
Toven retweeted
If you're a researcher looking to:
β conduct rigorous studies on how multiple models can outperform the frontier
β leverage data from the largest LLM marketplace (150 trillion tokens processed per month)
... DM me with your work!
We have an exciting role coming in the future, but might fill it opportunistically.
Introducing the Fusion API, the smartest compound model in the market.
Fusion achieves Fable-level intelligence at half the price.
How it works π
20
15
229
39,365
19h
hi @karpathy model council is now just an api call away
Introducing the Fusion API, the smartest compound model in the market.
Fusion achieves Fable-level intelligence at half the price.
How it works π
1
3
41
3,310
20h
we have fable at home
Introducing the Fusion API, the smartest compound model in the market.
Fusion achieves Fable-level intelligence at half the price.
How it works π
10
3
59
8,904
Toven retweeted
Introducing the Fusion API, the smartest compound model in the market.
Fusion achieves Fable-level intelligence at half the price.
How it works π
486
1,256
11,002
3,309,272
Toven retweeted
Jun 13
New server tool: Subagent π€
Your model can now delegate focused sub-tasks to a smaller, cheaper, faster model mid-generation.
The big model orchestrates, the subagent executes. The subagent can use any model on OpenRouter.
12
45
577
37,831
Jun 11
should be interesting!
Monday: Darkbloom goes live on OpenRouter.
Every OpenRouter user gets a free trial of private inference, powered by idle Macs around the world.
And for providers -- next week we're launching an Alpha Rewards Program: earn up to $40 guaranteed just for running a node. Limited time while we're in alpha. Breakdown in the replies.
The network is ready for it. Since our Public Alpha two weeks ago: Gemma 4 is now multimodal, SSD caching has made time-to-first-token much faster, and load tests show we can sustain millions of tokens per second -- with more upgrades landing before launch.
If you run a node: turn it on. Demand arrives Monday.
If you have a Mac: darkbloom(.)dev β one install command and you're earning.
The world's sleeping compute is waking up.
1
1
3
686
Toven retweeted
Jun 11
Use our Benchmarks explorer to plot Pareto curves for 10 different benchmarks, including @ArtificialAnlys and @Designarena:
openrouter.ai/rankings#benchβ¦
2
7
56
11,888
hi we desperately need a technical product marketing person at @OpenRouter
if things like
- thinking 6 months ahead
- making devs successful with content
- represent end users of openrouter rather than openrouter's comfort
- complete autonomy agency
pls apply ππ»
8
3
61
5,198
Jun 11
i am hir*ng for my team (provider operations)
i need inference api nerds. i need the nerd who undestands why the interactions api needed to happen. i need nerds who appreciate the responses api for the gaps it plugs in chat completions.
i need this niche ass nerd. help!
4
4
37
2,318
Jun 11
i need people who understand the difference between reasoning_effort and output_config.effort and verbosity and thinkingLevel and thinkingBudget and budget_tokens and max_tokens and max_completion_tokens
1
5
169
Jun 11
i need people who get it like @KTibow gets it but who have also felt the pain of implementation and can feel the limits, and have the intuition and creativity to build past those limits x.com/ktibow/status/20649051β¦
something beautiful in that openrouter hires those who understand responses/interactions so you can keep using chat completions as long as you like
1
252
Jun 11
1
4
257
no benchmark will tell you this: LLMs can be /too/ nice
unsurprisingly, in a competitive zero-sum setting, being nice can be bad
i built royale: last agent standing, a br for agents, and ran it 30 times
the nicest model lost hard. the model you least expected, won
π§΅:
10
9
50
17,658
Toven retweeted
Jun 10
Today we're launching the new Activity explorer on OpenRouter.
It's the best way to see how much and your team are spending on every model, along with tokens, cache hit rate, agents, & trends. All updated in real time.
See how our team is using Fable and other models π
11
12
198
32,511