DECISION & CLASSIFICATION / Together AI
Tev1
An experimental fine-tuned model for selecting one supplied option, with a public classifier-training recipe.
ACCESS & CAPABILITIES
How it fits your workflow
The repository links weights and a Together endpoint; endpoint access depends on the account. Hosted pricing is not verified here.
- Documented judgment types
- Choose an option
- Languages
- Language coverage is not verified in this catalog.
- Context and input limits
- The published training recipe uses a 2,048-token sequence limit and 2–24 answer options; serving limits require a separate check.
- Jev-compatible interface
- Different interface
PRICE & PROBABILITIES
Hosted price unverified
A hosted endpoint is documented, but its rate is unverified here. Self-hosting also has hardware and operating costs.
Interpret the output
Returns an option letter and token logprobs. The repository explicitly says these are model preferences, not calibrated confidence.
Check before choosing
Experimental, choice-focused interface; do not assume Jev's Score and Noul contract.
The headline repository results reuse development benchmarks, so they are not an untouched final test.
No shared benchmark measured yet
Our accuracy, calibration, p50/p95 latency and cost per thousand decisions are not measured. This is a source-reviewed capability profile. Compare the same task, model version, language, input size and serving setup before ranking results.
Read the selection and evaluation guide →Official sources & version notes
Together · Tev1 repository ↗ Checked 2026-10-05