meraGPT's hosted decision model (state-decider-1). Answers yes/no, choice and rubric-score questions over one state as calibrated distributions in a single pass, on the System One schema, so the typesafe-sdk works by changing its base URL.
Together AI's Jev-style classifier. A LoRA fine-tune of Qwen3.5-4B that reads a state, a question and 2 to 24 options and returns one option letter. Served on Together's platform; recipe published.
Decides
choice, score, noul, classify, route
choice, classify, route
Architecture
state-decider
tev
Fine-tuned from
—
qwen/qwen3.5-4b
License
proprietary
Unspecified — weights licence being finalised
Availability
Hosted API
Open weights + hosted API
Hosted by
meraGPT
Together AI
Input price
$0.030/MTok
$0.042/MTok
Decision accuracy
76.8%
88.0%
Calibration error
—
—
Valid action rate
—
—
Median latency
—
—
p95 latency
—
—
Evaluation suite
Figures are from each model’s manifest; accuracy and latency are what the publishers report, on their own suites and hardware. Add a third model.
Questions
What is the difference between state-decider-1 and tev1?
state-decider-1 is from meraGPT and tev1 from Together AI. state-decider-1 is only available as a hosted API; tev1 has open weights and a hosted API. Both answer choice, classify and route questions. Only state-decider-1 answers score and noul. state-decider-1 is licensed proprietary; tev1, other.
Which is more accurate, state-decider-1 or tev1?
They report on different suites — state-decider-1 76.8% on typed-decisions benchmark (400 cases, 2,000 decisions; teacher-ensemble references), single run, tev1 88.0% on Together development set (reused, not held out) — so the numbers do not rank them. Test both on your own labelled examples.
Which is cheaper, state-decider-1 or tev1?
state-decider-1: $0.03 / $0 per 1M. tev1: $0.042 / $0 per 1M, or free to self-host. Open weights cost nothing per call beyond your own hardware.
Can I run state-decider-1 or tev1 locally?
tev1 yes — systemone pull together-ai/tev1 downloads its weights. The other is only served as a hosted API.
typed-decisions benchmark (400 cases, 2,000 decisions; teacher-ensemble references), single run
