@samsja19: It was a pleasure to collaborate with the talented @mixedbreadai team to push a sota agentic search with our prime rl s…
Summary
Samsja19 highlights their collaboration with Mixedbread AI on Toast 1, a specialized agentic search model that achieves frontier quality at 12x faster speed and 1/10th the cost via reinforcement learning.
View Cached Full Text
Cached at: 08/14/26, 11:32 AM
It was a pleasure to collaborate with the talented @mixedbreadai team to push a sota agentic search with our prime rl stack
RL is truly beautiful.
Even the biggest model don’t adapt well to new harness and new tool out of the box, but with RL you can get the model to outperform extremely well at a faction of the cost
Everybody spending meaningfully amount of money on inference will eventually need to postrain
Mixedbread (@mixedbreadai): Introducing Toast 1, our first specialised search agent.
Toast 1 sets a new Pareto frontier for agentic search models.
Frontier search quality, across all domains, 12x faster, at 1/10th of the price.
Similar Articles
Introducing Toast 1
Mixedbread introduces Toast 1, a specialized search agent that matches frontier model quality while being up to 10x cheaper and 12x faster. It automates agentic search loops and achieves state-of-the-art results on benchmarks like OfficeQA Pro V2 and legal knowledge tasks.
@lu__jasper: Some early results from playing around with search on a subsampled version of OBLIQ-bench. Mixedbread's reranker is a b…
Early results from testing search on a subsampled OBLIQ-bench show that Mixedbread's reranker achieves strong MRR, sometimes outperforming GPT 5.5 on certain metrics with faster speed, though the benchmark remains challenging.
@samsja19: Very exciting work to bridge the gap between RL and mid/pretraining You can learn from your environment beyond the rewa…
A new method called ECHO bridges RL and pre-training by using next token prediction on tool call outputs to learn from the environment beyond reward signals, combining world modeling and agentic actions.
@samsja19: prime rl can now express and train multi agent systems, enabling usecase like adjentic judge, self play, user simulatio…
Prime RL now supports expressing and training multi-agent systems, enabling use cases like agentic judge, self-play, user simulation, and complex agent collaboration.
@Vtrivedy10: there's a very exciting future agent recipe for building intelligence too cheap to meter, applied towards extracting si…
The post outlines a future agent recipe for building scalable intelligence by fine-tuning efficient, specialized open models to surpass frontier performance on LLM-as-a-judge tasks, and applying this to extract signals from trace data for continual learning. LangChain Labs and FireworksAI release new work demonstrating this approach.