Is EXAONE-4.0-1.2B-JEV v0.3 (unofficial fine-tune of LGAI-EXAONE/EXAONE-4.0-1.2B, native option-label softmax, two reads averaged, self-hosted) state of the art?#337Best current rank: #337 in Decision models, as of Nov 22, 2026.
As of Nov 22, 2026, EXAONE-4.0-1.2B-JEV v0.3 (unofficial fine-tune of LGAI-EXAONE/EXAONE-4.0-1.2B, native option-label softmax, two reads averaged, self-hosted) is ranked #337 in Decision models (since Nov 22, 2026; sources Decision Index Full and JevBench; methodology v1.0). Ranked #337 in Decision models.
Facts and scores
EXAONE-4.0-1.2B-JEV v0.3 (unofficial fine-tune of LGAI-EXAONE/EXAONE-4.0-1.2B, native option-label softmax, two reads averaged, self-hosted) is an open-weights (EXAONE AI Model License 1.2 (non-commercial)) text model from Unknown; 1.3B parameters and no public price.
How it scores
- On JevBench price it scores 2.5%, #69 of 134 (top 52%); 2.4 pts behind verdict-small (Manavarya09, multilingual-e5-small 118M); level with Sage 1.3.0 (Levanto Labs) above and level with Aplomb 1 below.
- On Decision Index ECE it scores 7.9%, #259 of 420 (top 62%); 6.8 pts behind KnowLine 9B Gen2; level with JPT-4B above and Jev 1.13.0 (TypeSafe AI) is 0.1 pts below.
- On JevBench it scores 47.7 (CI 47.2–48.2), #90 of 134 (top 68%); 35.1 behind Perplexity Decider v1.1 27B (Qwen3.8-27B, noncausal decision readout); level with Open-Jev 2B (Zefan Cai) above and level with SPX-CD Pro (SurdAI, hosted /v1/systemone, Oct-4 checkpoint) below.
History
First result on isitsota: Oct 8, 2026. Best overall rank so far: #337 in Decision models, first held Nov 22, 2026.
Reliability
Index bins for EXAONE-4.0-1.2B-JEV v0.3 (unofficial fine-tune of LGAI-EXAONE/EXAONE-4.0-1.2B, native option-label softmax, two reads averaged, self-hosted). Confidence on the horizontal axis, accuracy on the vertical axis.
Median 148.5 ms, same GPU.
$0.0253 per 1,000 decisions.
Evidence
| Source | Score | Date | Type | Receipt |
|---|---|---|---|---|
| Decision Index Full EXAONE-4.0-1.2B-JEV v0.3 | 20.6 20.1–21.1 | 2026-11-21 | INDEP. | snap_decision-index_2026-11-21_release-v4|decision-index-0.4.1 |
| JevBench EXAONE-4.0-1.2B-JEV v0.3 (unofficial fine-tune of LGAI-EXAONE/EXAONE-4.0-1.2B, native option-label softmax, two reads averaged, self-hosted) | 47.7 47.2–48.2 | 2026-10-08 | INDEP. | snap_jevbench_2026-10-08_v1.6.1|jevbench-v1.6.1 |
Where it sits in the succession
Crown intervals for this model are still empty.
Where it is not the crown
| Category | This model | Rank | Leader |
|---|---|---|---|
| Decision models | #337 in this category | #337 | karpo-decision-31B |
Related pages
Is EXAONE-4.0-1.2B-JEV v0.3 (unofficial fine-tune of LGAI-EXAONE/EXAONE-4.0-1.2B, native option-label softmax, two reads averaged, self-hosted) SOTA?
Best current rank: #337 in Decision models, as of Nov 22, 2026.
What is EXAONE-4.0-1.2B-JEV v0.3 (unofficial fine-tune of LGAI-EXAONE/EXAONE-4.0-1.2B, native option-label softmax, two reads averaged, self-hosted) best at?
As of Nov 22, 2026, EXAONE-4.0-1.2B-JEV v0.3 (unofficial fine-tune of LGAI-EXAONE/EXAONE-4.0-1.2B, native option-label softmax, two reads averaged, self-hosted) is ranked #337 in Decision models (since Nov 22, 2026; sources Decision Index Full and JevBench).