Since Claude Sonnet 5.5 and others became GA in Kiro, I tried to summarize the current list of models
This page has been translated by machine translation. View original
Introduction
The Kiro changelog from 2026-10-21 announced the GA of Claude Fable 5.1, Claude Sonnet 5.5, and GPT-5.6 Terra・Luna.
At the time of writing, I compiled a list of models available in the Kiro Enterprise (Pro) license environment, along with their release dates and credit multipliers.
Models with the Experimental Label Removed
According to the changelog, Fable 5.1 had its Preview label removed, and Sonnet 5.5, Terra, and Luna had their Experimental labels removed.
What Are Kiro Excluded Models?
Kiro Excluded Models is a classification for models to which the indemnity under AWS Service Terms Section 50.10 does not apply to their output. According to the docs, models labeled as experimental / pre-release / preview / beta, as well as Kimi K3, fall under this classification. According to the changelog, the 4 models that became GA have been removed from this classification.
There are currently 7 Experimental models.
| Model | Release Date | Credit Multiplier |
|---|---|---|
| Claude Haiku 5.5 | 2026-10-23 | 0.06x |
| Kimi K3 | 2026-10-15 | 2.4x |
| Claude Opus 5.5 | 2026-09-28 | 2.0x |
| GPT-5.6 Sol | 2026-07-13 | 4.4x |
| DeepSeek 3.2 | 2026-02-10 | 0.25x |
| MiniMax M2.1 | 2026-02-10 | 0.15x |
| Qwen3 Coder Next | 2026-02-10 | 0.05x |
Kiro Model List
You can retrieve the list of models available in your environment with the following command.
kiro-cli chat --list-models
The kiro-cli 2.34.0 list does not have a column indicating Experimental status. Experimental models can be identified by the fact that their description begins with "Experimental preview."
claude-opus-5.5 2.00x credits Experimental preview of Claude Opus 5.5 model with 1M context window
claude-sonnet-5.5 1.30x credits Claude Sonnet 5.5 model with 1M context window
The following table lists the models marked as Active in the docs. Release dates are taken from the Available models page in the docs, and credit multipliers are taken from the CLI list and the Models page in the docs.
| Model | Release Date | Credit Multiplier |
|---|---|---|
| Claude Sonnet 5.5 | 2026-10-02 | 1.3x |
| Claude Fable 5.1 | 2026-09-16 | 6x |
| Claude Opus 5 | 2026-07-24 | 2.2x |
| GPT-5.6 Terra | 2026-07-13 | 2.2x |
| GPT-5.6 Luna | 2026-07-13 | 0.6x |
| Claude Sonnet 5 | 2026-06-30 | 1.3x |
| Claude Opus 4.8 | 2026-05-28 | 2.2x |
| Claude Opus 4.7 | 2026-04-16 | 2.2x |
| GLM-5 | 2026-03-31 | 0.5x |
| MiniMax M2.5 | 2026-03-18 | 0.25x |
| Claude Sonnet 4.6 | 2026-02-17 | 1.3x |
| Claude Opus 4.6 | 2026-02-05 | 2.2x |
| Claude Opus 4.5 | 2025-11-24 | 2.2x |
| Claude Sonnet 4.5 | 2025-09-29 | 1.3x |
| Auto | 2025-09-17 | 1.0x |
| Claude Haiku 4.5 | 2025-09-04 | 0.4x |
The multipliers for the 3 GPT-5.6 models and Haiku 5.5 vary depending on the number of tokens in the request. For GPT-5.6, the multipliers in the table apply up to 272K tokens, and above that the multipliers double (Sol 8.8x, Terra 4.4x, Luna 1.2x). For Haiku 5.5, the multiplier is 0.06x up to 100K tokens, and 0.3x above that.
Summary
In Kiro as of September 2025, the available choices were Sonnet 4.0 and Haiku 4.5, but as of October 2026, a wide variety of models are now available, including Claude's frontier model Fable 5.1, Opus 5.5 which is said to offer comparable performance, OpenAI's GPT-5.6, and Kimi K3.
In Kiro, "Auto" is said to select and use the optimal model for most use cases, but we recommend switching between models depending on the use case and as a means of obtaining a second opinion.
Additionally, if your organization's policies have restricted the use of models due to their Experimental status, we recommend removing Sonnet 5.5 and GPT-5.6 Terra・Luna, which became GA this time, from those restrictions so they can be evaluated.
