Evan
6785807633
Refs librefang/librefang#4842 — long-term replacement for the substring match that the OpenAI driver currently uses to decide how to handle `reasoning_content` on historical assistant turns. Three provider-specific behaviours that the driver must distinguish at wire time, now expressed as catalog metadata: * `strip` — DeepSeek R1 / deepseek-reasoner. The API rejects requests that carry reasoning_content on previous assistant messages. * `echo` — DeepSeek V4 Flash. Thinking mode is on by default and the API rejects multi-turn requests when assistant turns containing tool_calls don't echo back the original reasoning text. This is the bug surfaced in librefang/librefang#4842. * `empty_string` — Moonshot / Kimi K2 family. The field must be present (empty string) on tool_calls turns, with thinking disabled wire-side for multi-turn compatibility. * `none` (default) — most providers; field is omitted entirely. V4 Pro is intentionally NOT marked `echo` — librefang#4842 reports it working out-of-the-box; flip when there's an empirical reproducer. Marks affected models: providers/deepseek.toml deepseek-v4-flash → echo deepseek-reasoner → strip providers/moonshot.toml kimi-k2.6, kimi-k2.5, kimi-k2 → empty_string providers/kimi-coding.toml kimi-for-coding → empty_string providers/byteplus-coding.toml kimi-k2.5 → empty_string providers/novita.toml moonshotai/kimi-k2-thinking → empty_string Tooling: * schema.toml registers the field with the four enum options and a `none` default so existing TOML files keep parsing unchanged. * scripts/validate.py rejects unknown enum values; verified with a hand-crafted negative case (`reasoning_echo_policy = "bogus"` → validation fails with the expected message). * `python3 scripts/validate.py` passes (267 models). The librefang side that consumes this field will land in a follow-up PR — until then, registry consumers ignore the field via `#[serde(default)]` and the existing substring fallback continues to work, so this commit is safe to ship independently.
57 lines
1.3 KiB
TOML
57 lines
1.3 KiB
TOML
# Moonshot / Kimi — https://moonshot.ai
|
|
# Models: 3
|
|
|
|
[provider]
|
|
id = "moonshot"
|
|
display_name = "Moonshot (Kimi)"
|
|
api_key_env = "MOONSHOT_API_KEY"
|
|
base_url = "https://api.moonshot.ai/v1"
|
|
key_required = true
|
|
|
|
[[models]]
|
|
id = "kimi-k2.6"
|
|
display_name = "Kimi K2.6"
|
|
tier = "frontier"
|
|
context_window = 262144
|
|
max_output_tokens = 16384
|
|
input_cost_per_m = 0.74
|
|
output_cost_per_m = 3.49
|
|
supports_tools = true
|
|
supports_vision = true
|
|
supports_streaming = true
|
|
supports_thinking = true
|
|
# Kimi requires reasoning_content present (empty string) on assistant
|
|
# turns with tool_calls, with thinking disabled wire-side for multi-turn
|
|
# compatibility. See librefang/librefang openai driver.
|
|
reasoning_echo_policy = "empty_string"
|
|
aliases = ["kimi", "kimi-k2.6-0420"]
|
|
|
|
[[models]]
|
|
id = "kimi-k2.5"
|
|
display_name = "Kimi K2.5"
|
|
tier = "frontier"
|
|
context_window = 131072
|
|
max_output_tokens = 16384
|
|
input_cost_per_m = 0.44
|
|
output_cost_per_m = 2.0
|
|
supports_tools = true
|
|
supports_vision = true
|
|
supports_streaming = true
|
|
supports_thinking = true
|
|
reasoning_echo_policy = "empty_string"
|
|
aliases = ["kimi-k2.5-0711"]
|
|
|
|
[[models]]
|
|
id = "kimi-k2"
|
|
display_name = "Kimi K2"
|
|
tier = "frontier"
|
|
context_window = 131072
|
|
max_output_tokens = 16384
|
|
input_cost_per_m = 0.57
|
|
output_cost_per_m = 2.3
|
|
supports_tools = true
|
|
supports_vision = true
|
|
supports_streaming = true
|
|
reasoning_echo_policy = "empty_string"
|
|
aliases = []
|