Skip to content

feat: add new models in model capabilities manager #1309

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
paulclindo merged 1 commit into from
Jan 28, 2026
Merged

feat: add new models in model capabilities manager #1309

Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
36 changes: 36 additions & 0 deletions shinkai-bin/shinkai-node/src/managers/model_capabilities_manager.rs
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -183,6 +183,23 @@ impl ModelCapabilitiesManager {
model_type if model_type.starts_with("ministral-3") => {
vec![ModelCapability::TextInference, ModelCapability::ImageAnalysis]
}
model_type if model_type.starts_with("functiongemma") => vec![ModelCapability::TextInference],
model_type if model_type.starts_with("translategemma") => {
vec![ModelCapability::TextInference, ModelCapability::ImageAnalysis]
}
model_type if model_type.starts_with("glm-4.7-flash") => vec![ModelCapability::TextInference],
model_type if model_type.starts_with("nemotron-3-nano") => vec![ModelCapability::TextInference],
model_type if model_type.starts_with("qwen3-vl") => {
vec![
ModelCapability::TextInference,
ModelCapability::ImageAnalysis,
ModelCapability::VideoAnalysis,
]
}
model_type if model_type.starts_with("deepseek-ocr") => {
vec![ModelCapability::TextInference, ModelCapability::ImageAnalysis]
}
model_type if model_type.starts_with("granite4") => vec![ModelCapability::TextInference],
model_type if model_type.starts_with("llama3.2-vision") => {
vec![ModelCapability::TextInference, ModelCapability::ImageAnalysis]
}
Expand Down Expand Up @@ -903,8 +920,17 @@ impl ModelCapabilitiesManager {
model_type if model_type.starts_with("gemma3:12b") => 128_000,
model_type if model_type.starts_with("gemma3:27b") => 128_000,
model_type if model_type.starts_with("gemma3") => 128_000,
// qwen3-vl must precede qwen3 (vision+video, 256K) — ollama.com/library/qwen3-vl
model_type if model_type.starts_with("qwen3-vl") => 256_000,
model_type if model_type.starts_with("qwen3") => 32_000,
model_type if model_type.starts_with("magistral") => 39_000,
model_type if model_type.starts_with("functiongemma") => 32_000,
model_type if model_type.starts_with("translategemma") => 128_000,
model_type if model_type.starts_with("glm-4.7-flash") => 198_000,
model_type if model_type.starts_with("nemotron-3-nano") => 1_000_000,
model_type if model_type.starts_with("deepseek-ocr") => 8_192,
model_type if model_type.starts_with("granite4:350m") => 32_000,
model_type if model_type.starts_with("granite4") => 128_000,
_ => 4096, // Default token count if no specific model type matches
}
}
Expand Down Expand Up @@ -1165,6 +1191,11 @@ impl ModelCapabilitiesManager {
|| model.model_type.starts_with("mistral-large")
|| model.model_type.starts_with("mistral-pixtral")
|| model.model_type.starts_with("ministral-3")
|| model.model_type.starts_with("functiongemma")
|| model.model_type.starts_with("glm-4.7-flash")
|| model.model_type.starts_with("nemotron-3-nano")
|| model.model_type.starts_with("qwen3-vl")
|| model.model_type.starts_with("granite4")
|| model.model_type.starts_with("qwen2.5-coder")
|| model.model_type.starts_with("qwen2.5vl")
|| model.model_type.starts_with("qwq")
Expand Down Expand Up @@ -1221,6 +1252,11 @@ impl ModelCapabilitiesManager {
|| model.model_type.starts_with("mistral-large")
|| model.model_type.starts_with("mistral-pixtral")
|| model.model_type.starts_with("ministral-3")
|| model.model_type.starts_with("functiongemma")
|| model.model_type.starts_with("glm-4.7-flash")
|| model.model_type.starts_with("nemotron-3-nano")
|| model.model_type.starts_with("qwen3-vl")
|| model.model_type.starts_with("granite4")
|| model.model_type.starts_with("magistral")
|| model.model_type.starts_with("gpt-oss")
}
Expand Down