Repository navigation
lm15 2.3.0 for Rust
In Rust: base_models on RouterConfig (and on LmBuilder), then "tinker:tinker://…" as the model.
Tinker (Thinking Machines): base models and the models you trained there (lm15-contract 18b3788; 2,266 of 2,266 checks, the same in every lm15 language; the new rules MAP-21, MAP-22 and MAP-23 are decided and wait for ratification).
- Two new providers, one key (
TINKER_API_KEY).tinkerspeaks Chat Completions (also spelledthinkingmachines);tinker-anthropicspeaks the Anthropic Messages wire. Both serve Tinker's base models (tinker:Qwen/Qwen3.5-4B) and any sampler checkpoint by its path. - A URI is one model name.
tinker://<run>:train:0/sampler_weights/<n>routes to Tinker whole. Before, anyname://…string was read as providernamewith model//…. - Reasoning is switched the way each model family takes it. On
tinker, Qwen 3/3.5/3.6, Kimi K2.6 and Nemotron take an on/off switch (an effort word turns reasoning on, and the substitution is recorded); Inkling takes the effort words. A checkpoint follows the family of the model it was trained from when you name that model (the line above). Listing models ontinkershows your checkpoints as fine-tunes with their base model. - Error messages from FastAPI servers are kept. An error body that is
{"detail": "…"}used to raise with an empty message on every OpenAI-family and Anthropic-family adapter; the detail is now the message (MAP-21). stopmeans the same everywhere. Tinker returns the stop sequence it stopped at; lm15 cuts it off (MAP-22).- New compat values: thinking format
reasoning_effort_switch, reasoning offswitch, JSON schemanone(a server that applies noresponse_format: dropped with a record), andstop_in_text. Registry entries may carry aliases. - Known, not changed in this release: the reader for a Chat Completions response body that did not come through lm15 (
response_from_openai_chatin Python; the same door under each language's spelling) is given no request, so a Tinker body that stopped at a stop sequence keeps the sequence at its end. Ordinary calls and streams cut it. On the Messages door, Tinker itself refuses reasoning off and effort words for the on/off families; usetinkerfor those.
Nothing is removed and no field changes meaning. Tinker calls both of its doors beta.
Rust source compatibility. The chat core covered by the 2.x promise is unchanged. Outside it, this release adds public fields and enum variants, which cargo semver-checks reports as breaking for code that lists every field or every variant:
- new field
stop_in_textonOpenAIChatCompat,ResolvedOpenAIChatCompat,ChatModelOverride,AnthropicCompatandResolvedAnthropicCompat;aliasesonregistry::ProviderDefinition;base_modelonwire::BuildContext. A struct literal that names every field needs the new one (or..Default::default()where the type has a default). - new variants
OpenAIChatThinkingFormat::ReasoningEffortSwitch,OpenAIChatReasoningOff::SwitchandOpenAIChatJsonSchema::None. Amatchover one of these enums without a wildcard arm needs a new arm.