Lumma-Fev-27B-v2

Lumma-Fev-27B-v2 is a decision model from FrontiersMind, fine-tuned from Qwen3.8-27B. You give it a state (text) and typed questions; it returns a probability for every option in one forward pass. It never generates text, so there is nothing to parse. It works in English, Hindi, Tamil and Kannada (and Hinglish).

We used DecisionFormer for training our model.

What changed from Lumma-Fev-27B: - Bidirectional attention. The full-attention layers read the whole prompt in both directions, like an encoder; the Gated DeltaNet (linear-attention) layers stay causal. - About 300K training samples instead of about 100K.

Results (accuracy, same rows for every model)

Benchmark	Rows	Lumma-Fev-27B-v2	Lumma-Fev-27B	Jev
AG News	2,000	93.1	92.9	89.2
BANKING77	3,069	92.4	92.3	79.5
Emotion	1,989	91.7	91.0	58.6
RAGTruth	1,500	88.1	87.4	82.5
WinoGrande	1,000	92.1	82.6	90.7
Financial PhraseBank	999	79.3	81.5	77.7
JudgeBench	350	77.4	75.4	80.6
BBH	750	83.9	79.9	91.1
Average		87.3	85.4	81.2

None of these test rows appear in the training data. The model was trained on the train splits of AG News, BANKING77, Emotion, RAGTruth and WinoGrande. Financial PhraseBank, JudgeBench and BBH were not in its training data.

Calibration (average over the eight benchmarks, lower is better):

Metric	Lumma-Fev-27B-v2	Lumma-Fev-27B	Jev
Log loss	0.307	0.350	1.315
ECE	0.024	0.028	0.088
Brier	0.178	0.206	0.291

Run it with Ollama

ollama pull FrontiersMind/lumma-fev-27b-v2

Ask a yes/no question about a state from stdin:

ollama run FrontiersMind/lumma-fev-27b-v2 --ask "Does the customer need a reply today?" < ticket.txt

Pick one option, passing each option as name="description":

ollama run FrontiersMind/lumma-fev-27b-v2 \
  --ask "Which team should handle this request?" \
  --choice billing="Charges and refunds" \
  --choice support="Technical integration errors" \
  --choice sales="Questions about buying a product" \
  < ticket.txt

Python

Send a state and a set of named questions to lumma-fev-27b-v2

import ollama

response = ollama.systemone(
    model="FrontiersMind/lumma-fev-27b-v2",
    state="My integration keeps failing. Please help.",
    questions={
        "team": {
            "type": "choice",
            "instructions": "Which team should handle this request?",
            "criteria": {
                "billing": "Charges and refunds",
                "support": "Technical integration errors",
                "sales": "Questions about buying a product",
            },
        },
        "urgent": {"type": "noul", "instructions": "Does the customer need a reply today?"},
    },
)

print(response)

Supported Decision Types

Type	Criteria	Answer fields
`noul`	A yes/no question about the state	`noul` (probability of yes)
`choice`	Each option name mapped to a description	`choice`, `probabilities`, `confidence`
`score`	Ordered levels, lowest first	Accepted, but untested

Each question reads the state and its own options only, so one answer never changes another. A question plus the state must fit in 8,192 tokens. The model was trained on noul and choice questions; score questions are accepted but untested.

Hardware

The full-precision weights need about 55 GB of free GPU memory (an 80 GB card). The Ollama build is 17 GB.

Full details are in the model card.

License

Apache 2.0

Lumma-fev-27b-v2 is a decision model from Frontiersmind.ai, fine-tuned from Qwen3.8-27B.

Models

Readme