cmd/config: add minimax-m2.5 (#14223)

This commit is contained in:
Parth Sareen
2026-02-12 14:29:50 -08:00
committed by GitHub
parent 935a48ed1a
commit 598b74d42c
3 changed files with 37 additions and 35 deletions

View File

@@ -24,25 +24,6 @@ type cloudModelLimit struct {
Output int
}
// cloudModelLimits maps cloud model base names to their token limits.
// TODO(parthsareen): grab context/output limits from model info instead of hardcoding
var cloudModelLimits = map[string]cloudModelLimit{
"cogito-2.1:671b": {Context: 163_840, Output: 65_536},
"deepseek-v3.1:671b": {Context: 163_840, Output: 163_840},
"deepseek-v3.2": {Context: 163_840, Output: 65_536},
"glm-4.6": {Context: 202_752, Output: 131_072},
"glm-4.7": {Context: 202_752, Output: 131_072},
"gpt-oss:120b": {Context: 131_072, Output: 131_072},
"gpt-oss:20b": {Context: 131_072, Output: 131_072},
"kimi-k2:1t": {Context: 262_144, Output: 262_144},
"kimi-k2.5": {Context: 262_144, Output: 262_144},
"kimi-k2-thinking": {Context: 262_144, Output: 262_144},
"nemotron-3-nano:30b": {Context: 1_048_576, Output: 131_072},
"qwen3-coder:480b": {Context: 262_144, Output: 65_536},
"qwen3-coder-next": {Context: 262_144, Output: 32_768},
"qwen3-next:80b": {Context: 262_144, Output: 32_768},
}
// lookupCloudModelLimit returns the token limits for a cloud model.
// It tries the exact name first, then strips the ":cloud" suffix.
func lookupCloudModelLimit(name string) (cloudModelLimit, bool) {