A Context Compression tool for #AI Agents. Fewer input tokens. Lower costs. Longer runs. Built on #Bittensor SN114 by @DendriteHQ
Joined January 2026
- Tweets238
- Following33
- Followers989
- Likes410
All for one, one for all.
That was great to launch three new Bittensor integrations in the past month: @say_gm_, @engyai, @chutes_ai.
Try SOMA: app.thesoma.ai/register
Claude Code is now supported on SOMA.
Our latest harness integration joins Codex and Copilot, bringing SOMA to more of the tools you code with.
Try Claude Code with SOMA: app.thesoma.ai
SOMA retweeted
Two ways to cut your agent bill: send fewer tokens, or pay less for each one.
With @SomaSubnet and SayGM, you get both on the same request.
On DeepSeek V4.1 Flash:
28% saved by compression on SOMA
55% model discount on SayGM
Full case study below 👇
saygm.com/blog/how-soma-cuts…
The power of Bittensor ☀️
Two ways to cut your agent bill: send fewer tokens, or pay less for each one.
With @SomaSubnet and SayGM, you get both on the same request.
On DeepSeek V4.1 Flash:
28% saved by compression on SOMA
55% model discount on SayGM
Full case study below 👇
saygm.com/blog/how-soma-cuts…
Meanwhile, in our datacenter...
You can fit more into your @MistralAI context window and let it work longer for the same price and quality.
Now available for Mistral Large 4.
Register: app.thesoma.ai/register
We have the best context compressor around! 🧵 1/4
SOMA can provide 30% more effective compute to your system. This gain has been measured against an unchanged baseline, with identical tasks, model, provider, environment and generation settings. thesoma.ai/labs/experiment-m…
🧵 3/4
We've benchmarked SOMA across 6 models, with token savings ranging from 18.91% to 47.10%, while maintaining comparable or better solve rates.
We repeat runs because agent trajectories are stochastic: one run can be short while another loops, explores a dead end or hits provider noise.
Repeating runs lets us distinguish stable improvements from lucky runs.
🧵 3/4
We've benchmarked SOMA across 6 models, with token savings ranging from 18.91% to 47.10%, while maintaining comparable or better solve rates.
We repeat runs because agent trajectories are stochastic: one run can be short while another loops, explores a dead end or hits provider noise.
Repeating runs lets us distinguish stable improvements from lucky runs.
Why is SOMA such a great compressor in the #Bittensor ecosystem?
One reason is that it extends your context window by 20%, allowing you to run longer sessions without hitting limits.
Shout-out to @infinitetensor and @_overment for the inspiration!
SOMA Highlights - October 9
We integrated SOMA with new inference providers: @chutes_ai, @say_gm_, and @engyai, and started proxing traffic to Albedo.
We had a great time on @opentensor Novelty Search and @ExploitSummit.
Check out the latest updates below.
Compress. Pay less.
SOMA retweeted
Bittensor gets stronger when its subnets use each other’s products.
That’s why we support the Gamma token initiative @const_reborn presented at @exploitsummit. It could make it easier for subnets to use each other’s products, and we want to help make that happen.
We were proud to see @SomaSubnet used as an example. @oli_soma also heard plenty of positive feedback on SOMA and came back with ideas for possible integrations after talking with other subnet owners. Gamma could give us more ways to work together.
The summit gave us time to meet fellow builders, hear what they’re working on and share what we’ve learned ourselves.
Our founder @pbarbachowski shared lessons from years of mining Bittensor and becoming a subnet operator. If you missed his talk, you can watch it here: youtube.com/watch?v=rNggdiur…
@hux_dendrite and @hunter_tensor ran a workshop on getting started as a miner and thinking like one. A few attendees even skipped school to be there.
Spending time together as a team was another highlight. It was great to have @teutonic_tony and @albedo_wthrboss there and introduce more of our builders to the wider Bittensor community.
Thanks to everyone who stopped by our booth for a chat. You cleared us out of merch and stickers in no time.
Congratulations @admiralagrippa, @SiliconJose and the whole @exploitsummit team - and thank you for bringing the Bittensor community together.
If you have an idea you want to bring to Bittensor, let’s talk. Let’s build your subnet!
