feat(openai-go): add EINO model component using official OpenAI Go SDK

Create a new `components/model/openai-go` module implementing EINO model
support backed by `github.com/openai/openai-go/v3` and the Responses
API. The component supports tool calling (including required/choices),
multimodal inputs, and streaming.

Co-authored-by: Octium Agent <agent@octium.dev>

Author: mikeb26

components/model/openai-go/README.md

@@ -0,0 +1,80 @@
+# OpenAI (official openai-go SDK)
+
+An OpenAI model implementation for [Eino](https://github.com/cloudwego/eino) using the official OpenAI Go SDK (`github.com/openai/openai-go/v3`). This is intended as a starting point for eventual replacement of the existing openai implementation (see ../openai) which is based on github.com/sashabaranov/go-openai. Newer models from OpenAI increasingly do not fully support the older chat completions API which github.com/sashabaranov/go-openai is based on. Consequently, this component targets the **Responses API only**.
+
+## Features
+
+- Implements `github.com/cloudwego/eino/components/model.ToolCallingChatModel`
+- Responses API (non-stream + streaming)
+- Tool calling support (function tools)
+- Multimodal inputs via `schema.Message.UserInputMultiContent`:
+  - text
+  - image_url (URL or base64 via `Base64Data` + `MIMEType`)
+  - file_url (URL or base64)
+
+## Installation
+
+```bash
+go get github.com/cloudwego/eino-ext/components/model/openai-go@latest
+```
+
+## Quick start
+
+```go
+package main
+
+import (
+  "context"
+  "log"
+  "os"
+
+  "github.com/cloudwego/eino/schema"
+  "github.com/cloudwego/eino-ext/components/model/openai-go"
+)
+
+func main() {
+  ctx := context.Background()
+
+  cm, err := openaigo.NewChatModel(ctx, &openaigo.Config{
+    APIKey: os.Getenv("OPENAI_API_KEY"),
+    Model:  "gpt-5.4", // any Responses API capable model
+  })
+  if err != nil {
+    log.Fatal(err)
+  }
+
+  out, err := cm.Generate(ctx, []*schema.Message{
+    {Role: schema.User, Content: "Hello"},
+  })
+  if err != nil {
+    log.Fatal(err)
+  }
+
+  log.Println(out.Content)
+}
+```
+
+## Tool calling
+
+Bind tools using `WithTools()`:
+
+```go
+cm2, err := cm.WithTools([]*schema.ToolInfo{
+  {
+    Name: "get_weather",
+    Desc: "Get weather at the given location",
+    ParamsOneOf: schema.NewParamsOneOfByParams(map[string]*schema.ParameterInfo{
+      "location": {Type: schema.String, Required: true},
+    }),
+  },
+})
+```
+
+Then control selection with Eino common options:
+
+- `model.WithTools(...)`
+- `model.WithToolChoice(schema.ToolChoiceAllowed|Forced|Forbidden, allowedToolNames...)`
+
+## Streaming
+
+Use `Stream()` to receive incremental `*schema.Message` deltas.

components/model/openai-go/chatmodel.go

@@ -0,0 +1,189 @@
+/*
+ * Copyright 2026 CloudWeGo Authors
+ *
+ * Licensed under the Apache License, Version 2.0 (the "License");
+ * you may not use this file except in compliance with the License.
+ * You may obtain a copy of the License at
+ *
+ *     http://www.apache.org/licenses/LICENSE-2.0
+ *
+ * Unless required by applicable law or agreed to in writing, software
+ * distributed under the License is distributed on an "AS IS" BASIS,
+ * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied.
+ * See the License for the specific language governing permissions and
+ * limitations under the License.
+ */
+
+package openaigo
+
+import (
+	"context"
+	"errors"
+	"fmt"
+	"net/http"
+	"time"
+
+	"github.com/cloudwego/eino/callbacks"
+	"github.com/cloudwego/eino/components"
+	"github.com/cloudwego/eino/components/model"
+	"github.com/cloudwego/eino/schema"
+	"github.com/openai/openai-go/v3"
+	"github.com/openai/openai-go/v3/option"
+	"github.com/openai/openai-go/v3/responses"
+)
+
+var _ model.ToolCallingChatModel = (*ChatModel)(nil)
+
+type Config struct {
+	APIKey string `json:"api_key"`
+
+	// Timeout specifies the maximum duration to wait for API responses.
+	// If HTTPClient is set, Timeout will not be used.
+	// Optional. Default: no timeout
+	Timeout time.Duration `json:"timeout"`
+
+	// HTTPClient specifies the client to send HTTP requests.
+	// If HTTPClient is set, Timeout will not be used.
+	// Optional. Default &http.Client{Timeout: Timeout}
+	HTTPClient *http.Client `json:"http_client"`
+
+	// BaseURL specifies the OpenAI endpoint URL
+	// Optional. Default: https://api.openai.com/v1
+	BaseURL string `json:"base_url"`
+
+	// Model specifies the ID of the model to use.
+	// Optional.
+	Model string `json:"model,omitempty"`
+
+	// MaxOutputTokens is an upper bound for the number of tokens that can be generated for a response,
+	// including visible output tokens and reasoning tokens.
+	MaxOutputTokens *int `json:"max_output_tokens,omitempty"`
+
+	TopP        *float32 `json:"top_p,omitempty"`
+	Temperature *float32 `json:"temperature,omitempty"`
+
+	// Reasoning config for reasoning models.
+	Reasoning *Reasoning `json:"reasoning,omitempty"`
+
+	// Store indicates whether to store the generated model response for later retrieval.
+	Store *bool `json:"store,omitempty"`
+
+	// Metadata set of key-value pairs that can be attached to an object.
+	Metadata map[string]string `json:"metadata,omitempty"`
+
+	// ExtraFields will override any existing fields with the same key.
+	// Optional. Useful for experimental features not yet officially supported.
+	ExtraFields map[string]any `json:"extra_fields,omitempty"`
+}
+
+type ChatModel struct {
+	cli openai.Client
+
+	model       string
+	maxOutTok   *int
+	topP        *float32
+	temperature *float32
+	reasoning   *Reasoning
+	store       *bool
+	metadata    map[string]string
+	extraFields map[string]any
+
+	tools      []responses.ToolUnionParam
+	rawTools   []*schema.ToolInfo
+	toolChoice *schema.ToolChoice
+}
+
+func NewChatModel(_ context.Context, config *Config) (*ChatModel, error) {
+	if config == nil {
+		return nil, fmt.Errorf("config cannot be nil")
+	}
+
+	opts := make([]option.RequestOption, 0, 4)
+	if config.APIKey != "" {
+		opts = append(opts, option.WithAPIKey(config.APIKey))
+	}
+	if config.BaseURL != "" {
+		opts = append(opts, option.WithBaseURL(config.BaseURL))
+	}
+	if config.HTTPClient != nil {
+		opts = append(opts, option.WithHTTPClient(config.HTTPClient))
+	} else if config.Timeout > 0 {
+		opts = append(opts, option.WithHTTPClient(&http.Client{Timeout: config.Timeout}))
+	}
+
+	cli := openai.NewClient(opts...)
+
+	cm := &ChatModel{
+		cli:         cli,
+		model:       config.Model,
+		maxOutTok:   config.MaxOutputTokens,
+		topP:        config.TopP,
+		temperature: config.Temperature,
+		reasoning:   config.Reasoning,
+		store:       config.Store,
+		metadata:    cloneStringMap(config.Metadata),
+		extraFields: cloneAnyMap(config.ExtraFields),
+	}
+
+	return cm, nil
+}
+
+func (cm *ChatModel) Generate(ctx context.Context, in []*schema.Message, opts ...model.Option) (outMsg *schema.Message, err error) {
+	ctx = callbacks.EnsureRunInfo(ctx, cm.GetType(), components.ComponentOfChatModel)
+
+	params, cbIn, err := cm.buildParams(in, false, opts...)
+	if err != nil {
+		return nil, err
+	}
+
+	ctx = callbacks.OnStart(ctx, cbIn)
+	defer func() {
+		if err != nil {
+			callbacks.OnError(ctx, err)
+		}
+	}()
+
+	resp, err := cm.cli.Responses.New(ctx, params)
+	if err != nil {
+		return nil, err
+	}
+
+	outMsg, err = cm.convertResponseToMessage(resp)
+	if err != nil {
+		return nil, err
+	}
+
+	callbacks.OnEnd(ctx, &model.CallbackOutput{
+		Message:    outMsg,
+		Config:     cbIn.Config,
+		TokenUsage: toModelTokenUsage(outMsg.ResponseMeta),
+		Extra: map[string]any{
+			callbackExtraModelName: string(resp.Model),
+		},
+	})
+
+	return outMsg, nil
+}
+
+func (cm *ChatModel) WithTools(tools []*schema.ToolInfo) (model.ToolCallingChatModel, error) {
+	if len(tools) == 0 {
+		return nil, errors.New("no tools to bind")
+	}
+	openAITools, rawTools, err := toOpenAITools(tools)
+	if err != nil {
+		return nil, err
+	}
+
+	tc := schema.ToolChoiceAllowed
+	ncm := *cm
+	ncm.tools = openAITools
+	ncm.rawTools = rawTools
+	ncm.toolChoice = &tc
+	return &ncm, nil
+}
+
+const typ = "OpenAI"
+
+func (cm *ChatModel) GetType() string { return typ }
+
+func (cm *ChatModel) IsCallbacksEnabled() bool { return true }