Tasks and inputs
Claude Haiku 5.5 is listed for Image-text-to-text, Text generation. It accepts text, image, pdf and produces text. Use the publisher’s input format and preprocessing for your chosen deployment.
SyncDev model directory · Profile updated Oct 13, 2026
Claude’s small model for high-volume extraction, classification and routing, with adaptive thinking and a 1M-token context window.
Pricing conditions: Starting rates apply to prompts of 100K input tokens or fewer. Longer prompts cost $0.50/M input and $2.50/M output. Token counts change with the newer tokenizer.
Recorded modality support shows what this model can accept and produce. An undocumented modality is shown as unknown rather than assumed to be unsupported.
Accepts
Produces
Claude Haiku 5.5 is a candidate for the repetitive work inside an application: extracting fields, assigning labels, deciding which workflow to run, and handling narrowly scoped assistant tasks. Anthropic released it on 7 October 2026. It accepts text and images and returns text. Its 1M-token context window and 128K-token ordinary output limit give it room for large inputs, but neither limit establishes the accuracy of a particular extraction or classification.
The model supports adaptive thinking controlled through effort. Start a migration with a small set of real requests and inspect the returned fields, rather than copying sampling settings from an older model. In particular, check Anthropic’s migration guide for unsupported temperature and sampling parameters. Thinking blocks are tied to the producing account or a linked account, which matters when a workflow moves requests between accounts.
The first-party starting rates are $0.10 per million input tokens and $0.50 per million output tokens for prompts of up to 100K input tokens. Above that threshold the documented rates are $0.50 input and $2.50 output. The starting number on a specification table is therefore not a flat rate for the entire context window. Check cache and batch terms separately.
Haiku 5.5 uses a newer tokenizer than Haiku 4.5. Recount the same documents before comparing bills; a cheaper token does not imply that a document uses the same number of tokens.
For extraction, include missing fields, conflicting values and documents that should be rejected. For routing, compare the selected destination with an agreed label and include an explicit fallback. Measure the cost of accepted results after retries. Choose a larger model only if the errors in that test justify its additional cost. These are evaluation suggestions; SyncDev has not independently benchmarked this model.
The overview, migration guide and pricing page below are the references for this profile. Compare Haiku 5.5 with Sonnet 5.5 when the same workflow needs a stronger general assistant, or browse Anthropic models.
Compare the model on the work you need to complete.
Claude Haiku 5.5 is listed for Image-text-to-text, Text generation. It accepts text, image, pdf and produces text. Use the publisher’s input format and preprocessing for your chosen deployment.
Documented limits: 1M context and 128K maximum output. The listed route has starting rates of $0.10 input and $0.50 output per million tokens. Measure a completed task using the prompts, documents and tools your application will use. Include reasoning and retries in the result.
Use the model identifier and account requirements for the selected API route. Confirm its request limits and output format with one complete task before connecting a larger workflow.
These records share a directory category. Check the source card, output type and deployment requirements before treating a model as a replacement.
Mistral AI
Mistral Large 4 is an open-weight multimodal model released in public preview on the publisher’s announced release, for general reasoning, coding and agent applications.
OpenAI
GPT-6.1 Sol is OpenAI's updated Sol model for coding and professional tasks. It accepts text and images and returns text.
inclusionAI
Ling 3.1 Flash is inclusionAI’s hybrid reasoning model for coding and tool agents, available through a hosted AI Gateway route.
Anthropic
Claude Sonnet 5.5 is an Anthropic language model that accepts text and images and returns text.
MiniMax
MiniMax M3.1 Flash Preview is a multimodal reasoning model for code and agent workflows, accessed through eligible MiniMax plans and tools.
Meituan
LongCat-2.5-Preview from Meituan: Meituan's multimodal reasoning model for coding and agent tasks, with image understanding and a 1M-token context window.