Skip to content
Open
Show file tree
Hide file tree
Changes from all commits
Commits
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
5 changes: 4 additions & 1 deletion README.md
Original file line number Diff line number Diff line change
Expand Up @@ -1751,7 +1751,7 @@ PROXY_URL=http://your-proxy:8080

#### Supported Models

PentAGI supports 21 AWS Bedrock models with tool calling, streaming, and multimodal capabilities. Models marked with `*` are used in default configuration.
PentAGI supports 22 AWS Bedrock models with tool calling, streaming, and multimodal capabilities. Models marked with `*` are used in default configuration.

| Model ID | Provider | Thinking | Multimodal | Price (Input/Output) | Use Case |
| ------------------------------------------------ | --------------- | -------- | ---------- | -------------------- | --------------------------------------- |
Expand All @@ -1760,6 +1760,7 @@ PentAGI supports 21 AWS Bedrock models with tool calling, streaming, and multimo
| `us.amazon.nova-pro-v1:0` | Amazon Nova | ❌ | ✅ | $0.80/$3.20 | Balanced accuracy, speed, cost |
| `us.amazon.nova-lite-v1:0` | Amazon Nova | ❌ | ✅ | $0.06/$0.24 | Fast processing, high-volume operations |
| `us.amazon.nova-micro-v1:0` | Amazon Nova | ❌ | ❌ | $0.035/$0.14 | Ultra-low latency, real-time monitoring |
| `us.anthropic.claude-opus-4-7` | Anthropic | ✅ | ✅ | $5.00/$25.00 | Latest Opus with adaptive thinking |
| `us.anthropic.claude-opus-4-6-v1`* | Anthropic | ✅ | ✅ | $5.00/$25.00 | World-class coding, enterprise agents |
| `us.anthropic.claude-sonnet-4-6` | Anthropic | ✅ | ✅ | $3.00/$15.00 | Frontier intelligence, enterprise scale |
| `us.anthropic.claude-opus-4-5-20251101-v1:0` | Anthropic | ✅ | ✅ | $5.00/$25.00 | Multi-day software development |
Expand All @@ -1780,6 +1781,8 @@ PentAGI supports 21 AWS Bedrock models with tool calling, streaming, and multimo

**Prices**: Per 1M tokens. Models with thinking/reasoning support additional compute costs during reasoning phase.

**Reasoning Modes**: Claude Opus 4.7 on Bedrock requires Adaptive mode, and Claude Opus/Sonnet 4.6+ support it. In provider settings, select `Reasoning Mode: Adaptive` and use `Reasoning Effort` to send Bedrock `output_config.effort` instead of a fixed thinking token budget.

#### Tested but Incompatible Models

Some AWS Bedrock models were tested but are **not supported** due to technical limitations:
Expand Down
9 changes: 8 additions & 1 deletion backend/pkg/database/converter/converter.go
Original file line number Diff line number Diff line change
Expand Up @@ -613,9 +613,13 @@ func ConvertAgentConfigToGqlModel(ac *pconfig.AgentConfig) *model.AgentConfig {
result.PresencePenalty = &ac.PresencePenalty
}

if ac.Reasoning.Effort != llms.ReasoningNone || ac.Reasoning.MaxTokens != 0 {
if !ac.Reasoning.IsZero() {
reasoning := &model.ReasoningConfig{}

if ac.Reasoning.Mode != pconfig.ReasoningModeDefault {
mode := model.ReasoningMode(ac.Reasoning.Mode)
reasoning.Mode = &mode
}
if ac.Reasoning.Effort != llms.ReasoningNone {
effort := model.ReasoningEffort(ac.Reasoning.Effort)
reasoning.Effort = &effort
Expand Down Expand Up @@ -708,6 +712,9 @@ func ConvertAgentConfigFromGqlModel(ac *model.AgentConfig) *pconfig.AgentConfig

if ac.Reasoning != nil {
reasoning := map[string]any{}
if ac.Reasoning.Mode != nil {
reasoning["mode"] = pconfig.ReasoningMode(*ac.Reasoning.Mode)
}
if ac.Reasoning.Effort != nil {
reasoning["effort"] = llms.ReasoningEffort(*ac.Reasoning.Effort)
}
Expand Down
78 changes: 77 additions & 1 deletion backend/pkg/graph/generated.go

Some generated files are not rendered by default. Learn more about how customized files appear on GitHub.

48 changes: 47 additions & 1 deletion backend/pkg/graph/model/models_gen.go

Some generated files are not rendered by default. Learn more about how customized files appear on GitHub.

12 changes: 11 additions & 1 deletion backend/pkg/graph/schema.graphqls
Original file line number Diff line number Diff line change
Expand Up @@ -23,13 +23,21 @@ enum ProviderType {
qwen
}

# Reasoning effort levels for advanced AI models (OpenAI format)
# Reasoning effort levels for advanced AI models
enum ReasoningEffort {
xhigh
max
high
medium
low
}

# Reasoning control mode for provider-specific thinking APIs
enum ReasoningMode {
adaptive
budget
}

# Template types for AI agent prompts and system operations
enum PromptType {
primary_agent
Expand Down Expand Up @@ -699,6 +707,7 @@ type ProviderConfig {

# AI model reasoning configuration
type ReasoningConfig {
mode: ReasoningMode
effort: ReasoningEffort
maxTokens: Int
}
Expand Down Expand Up @@ -748,6 +757,7 @@ type AgentsConfig {

# Input type for ReasoningConfig
input ReasoningConfigInput {
mode: ReasoningMode
effort: ReasoningEffort
maxTokens: Int
}
Expand Down
Loading