6 days ago
-
Document Seb-9B public JevBench and local Game Bench evaluations
-
Card: declare base model Qwen/Qwen3.5-9B; serve examples at 32k context (long benchmark states)
19 days ago
-
Card: all-rounder results on Laya public suites
-
Card: Ollama image note
-
Card: remove Decision Index section
-
Card: tagline
-
Remove unused client scripts
-
Model card: results, calibration, speed, GGUF/Ollama usage
-
model card: standard chat API, no custom client
-
model card: name Seb-9B
-
model card: Sol image reference scores
-
remove adapter and build records
-
Add files using upload-large-folder tool
20 days ago
-
model card: Mac latency on MLX, MLX serving
-
model card: Sev-9B
-
model card: name Seb-9B
-
model card: remove data collection details
-
model card: image-rejection claim limited to what we tested
-
model card: add Where r4 has the edge
-
initial commit
