add moonshot/kimi-k2.6 to model registry

[ishaan-berri]

Relevant Issues / Related PRs

Adds Kimi K2.6 (released April 20, 2026) to the LiteLLM model registry.

Pre-Submission Checklist

• I have added tests in tests/litellm/
• make test-unit passes for the modified tests

Changes

• Added moonshot/kimi-k2.6 to model_prices_and_context_window.json and backup
• Pricing: $0.60/M input, $2.80/M output (from https://platform.kimi.ai/docs/guide/kimi-k2-6-quickstart)
• 262K context window, supports function calling, vision, and video input
• Tests for model registry entry

Usage

import litellm

response = litellm.completion(
    model="moonshot/kimi-k2.6",
    messages=[{"role": "user", "content": "Hello!"}],
)

[CLAassistant]

Thank you for your submission! We really appreciate it. Like many open source projects, we ask that you sign our Contributor License Agreement before we can accept your contribution.
_{You have signed the CLA already but the status is still pending? Let us recheck it.}

[greptile-apps]

Greptile Summary

This PR registers moonshot/kimi-k2.6 in both the primary and backup model price/context-window JSON files, and adds a unit test suite that reads exclusively from the local backup (no network calls). The JSON pricing was verified against the official Kimi pricing page and is correct. Note: the PR description quotes different prices than what is in the JSON — the JSON is correct; the description appears to be a stale draft and should be updated.

Confidence Score: 5/5

Safe to merge — only adds a new model registry entry with verified pricing; no existing behavior changed.

All findings are P2. The JSON pricing values are correct per official Kimi docs. Both JSON files are in sync, the test suite makes no network calls, and all previously flagged missing fields (cache_read_input_token_cost, supports_reasoning) have been addressed. The only issue is a stale price in the PR description.

No files require special attention.

Important Files Changed

Filename	Overview
model_prices_and_context_window.json	Adds moonshot/kimi-k2.6 entry with pricing verified correct against official Kimi docs ($0.95/M input, $4.00/M output, $0.16/M cache-hit); all capability flags included.
litellm/model_prices_and_context_window_backup.json	Backup JSON kept in sync with the primary; identical moonshot/kimi-k2.6 entry added at the correct alphabetical position.
tests/test_litellm/llms/moonshot/test_moonshot_chat_transformation.py	New TestKimiK26ModelRegistry class validates presence, pricing, context window, capabilities, and provider; uses GetModelCostMap.load_local_model_cost_map() — no network calls, satisfying the unit-test-only rule.

Entity Relationship Diagram

%%{init: {'theme': 'neutral'}}%%
erDiagram
    MODEL_REGISTRY {
        string key "moonshot/kimi-k2.6"
        float input_cost_per_token "9.5e-07 ($0.95/M)"
        float output_cost_per_token "4e-06 ($4.00/M)"
        float cache_read_input_token_cost "1.6e-07 ($0.16/M)"
        int max_input_tokens "262144"
        int max_output_tokens "262144"
        int max_tokens "262144"
        string mode "chat"
        string litellm_provider "moonshot"
        boolean supports_function_calling "true"
        boolean supports_tool_choice "true"
        boolean supports_vision "true"
        boolean supports_video_input "true"
        boolean supports_reasoning "true"
    }
    MODEL_REGISTRY ||--|| BACKUP_JSON : "mirrored in"
    MODEL_REGISTRY ||--|| PRIMARY_JSON : "registered in"

Loading

_{Reviews (4): Last reviewed commit: "fix: load kimi-k2.6 registry tests from ..." | Re-trigger Greptile}

[greptile-apps]

Missing cache_read_input_token_cost vs kimi-k2.5

kimi-k2.5 (the predecessor) includes "cache_read_input_token_cost": 1e-07. If kimi-k2.6 also supports prompt caching (which is common for the Kimi K2 family), omitting this field means users won't benefit from cache-read cost savings when LiteLLM calculates cost. If cache reads are truly unsupported, a comment or zero-value entry would clarify intent.

[greptile-apps]

Missing supports_reasoning flag vs kimi-k2.5

kimi-k2.5 is registered with "supports_reasoning": true. If kimi-k2.6 is also a reasoning/thinking model, this field should be included so that LiteLLM's supports_reasoning() helper and related routing logic recognize the capability correctly without a future patch.

[gitguardian]

️✅ There are no secrets present in this pull request anymore.

If these secrets were true positive and are still valid, we highly recommend you to revoke them.
While these secrets were previously flagged, we no longer have a reference to the
specific commits where they were detected. Once a secret has been leaked into a git
repository, you should consider it compromised, even if it was deleted immediately.
Find here more information about risks.

---

^{_{🦉 GitGuardian detects secrets in your source code to help developers and security teams secure the modern development process. You are seeing this because you or someone else with access to this repository has authorized GitGuardian to scan your pull request.}}

[mateo-berri]

Could you fix the failing unit tests?

=========================== short test summary info ============================
FAILED tests/test_litellm/llms/moonshot/test_moonshot_chat_transformation.py::TestKimiK26ModelRegistry::test_kimi_k26_capabilities - KeyError: 'moonshot/kimi-k2.6'
FAILED tests/test_litellm/llms/moonshot/test_moonshot_chat_transformation.py::TestKimiK26ModelRegistry::test_kimi_k26_context_window - KeyError: 'moonshot/kimi-k2.6'
FAILED tests/test_litellm/llms/moonshot/test_moonshot_chat_transformation.py::TestKimiK26ModelRegistry::test_kimi_k26_in_model_cost_map - AssertionError: moonshot/kimi-k2.6 not found in model_cost
assert None is not None
FAILED tests/test_litellm/llms/moonshot/test_moonshot_chat_transformation.py::TestKimiK26ModelRegistry::test_kimi_k26_pricing - KeyError: 'moonshot/kimi-k2.6'
FAILED tests/test_litellm/llms/moonshot/test_moonshot_chat_transformation.py::TestKimiK26ModelRegistry::test_kimi_k26_provider - KeyError: 'moonshot/kimi-k2.6'
= 5 failed, 3466 passed, 73 skipped, 67 warnings, 10 rerun in 226.96s (0:03:46) =

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.

📢 Thoughts on this report? Let us know!

[mateo-berri]

LGTM. Awesome!