TMFusion Code 1.0
AVAILABLECost-efficient coding for software agents.
Directional ratings for model selection, not benchmark scores. More stars indicate broader capability, faster expected completion, or lower standard token pricing. Actual performance varies by task.
- Best fit
- Software engineeringImplementation · Debugging · Terminal tasks
- Context
- 512K tokensContext window
- Maximum output
- 128K tokensMaximum response length
- Standard price
- $2 · $6Input · output per 1M tokens
Fusion Code 1.0 combines strong coding capabilities with low token pricing for implementation, debugging, and terminal-based engineering tasks. It supports tool calling and structured outputs for software agents that inspect code, make changes, and test results. Text and image input are available through the Chat Completions and Responses APIs.
Use the Chat Completions or Responses API to connect Fusion Code 1.0 to your application.
- Model ID
- fusion-code-1.0
- API mode
- Stateless
- Text
- Input and output
- Image
- Input only
- Reasoning summaries
- Supported
MODEL GUIDE
Compare Fusion models
Compare model use cases, context limits, and standard token prices.
| Model | Status | Best fit | Context | Max output | Input / 1M | Output / 1M |
|---|---|---|---|---|---|---|
| Fusion Max 2.0 | Available | Coding, agents, and documents | 1M | 128K | $3.00 | $20.00 |
| Fusion Max 1.0 | Available | Reasoning and agents | 1M | 128K | $3.00 | $20.00 |
| Fusion Code 1.0 | Available | Software engineering | 512K | 128K | $2.00 | $6.00 |
Prices are in USD per million tokens at standard context rates. See each model's pricing details for cached-input and long-context rates.
Pricing
USD per million tokens, excluding applicable taxes.
| Context tier | Input | Cached input | Output |
|---|---|---|---|
| Standard, prompt ≤ 200K | $2.00 | $0.50 | $6.00 |
| Long context, prompt > 200K | $4.00 | $1.00 | $12.00 |
When input exceeds 200K tokens, long-context rates apply to the entire request. Tokens generated through internal multi-model collaboration are not billed. Usage is reported by the API.
Modalities
Fusion Code 1.0 accepts text and image input and returns text or tool calls.
- Text
- Input and output
- Instructions, conversation, and generated responses
- Image
- Input only
- Screenshots, diagrams, and visual context
- Audio
- Not supported
- Audio input and generation are not available
- Files and video
- Not supported directly
- Use agent tools to read files and pass their content to the model
Endpoints
Choose the endpoint supported by your application or agent client.
- Responses
- Supported
- POST /api/fusion/v1/responses
- Chat Completions
- Supported
- POST /api/fusion/v1/chat/completions
- Anthropic Messages
- Adapter required
- No native endpoint
- Embeddings, image, audio
- Not supported
- Not provided by these model endpoints
Endpoints are stateless. Send the complete conversation and tool history with each request.
Features
Supported features for application and agent integrations.
- Streaming
- Supported
- Server-sent events
- Function calling
- Supported
- JSON Schema tool arguments
- Custom tools
- Supported
- Freeform tool inputs via Responses
- Structured output
- Supported
- JSON object and JSON Schema
- Reasoning summaries
- Supported
- Auto, concise, or detailed
- Prompt caching
- Automatic
- No separate cache-write charge
- Fine-tuning
- Not supported
- No customer fine-tuning endpoint
Build with Fusion Code 1.0.
Get API access for your coding and software-agent workflows.
Request access