v0.8.10

What's Changed

• fix: use platform-appropriate model cache defaults by @leehack in #248
• docs: restore model cache release note history by @leehack in #250
• chore: prepare 0.8.10 release by @leehack in #249

Full Changelog: v0.8.9...v0.8.10

v0.8.9

What's Changed

• chore: restore pub score compatibility by @leehack in #247

Full Changelog: v0.8.8...v0.8.9

v0.8.8

What's Changed

• Add cross-platform model cache strategies by @leehack in #242
• chore(native): sync native release b9803 by @github-actions[bot] in #244
• Prepare 0.8.8 release by @leehack in #245
• docs: bump llama.cpp Flutter companion snippet by @leehack in #246

Full Changelog: v0.8.7...v0.8.8

v0.8.7

0.8.7

• Fixed multimodal chat-template rendering so templates that force-open reasoning, for example Qwen3.5 VLM prompts ending with <think>, preserve enable_thinking and stream generated reasoning through delta.thinking instead of delta.content.

Validation highlights:

• dart pub publish --dry-run
• GitHub CI on PR #240
• Local real-model smokes: Qwen3.5 multimodal macOS repro and Gemma 4 GGUF chat features smoke
• Tag-triggered Publish to pub.dev workflow
• Tag-triggered Docs Version Cut workflow
• GitHub Pages docs deploy
• ./tool/docs/validate_links.sh

v0.8.6

0.8.6

Patch release for the llama.cpp native runtime sync to leehack/llamadart-native@b9776.

• Updated the default llama.cpp native runtime pin to leehack/llamadart-native@b9776.
• Regenerated the matching Dart FFI bindings.
• Refreshed the llamadart_llama_cpp_flutter Apple SwiftPM checksum.
• Aligned README and website native override docs for the new native pin and companion install snippets.

Validation highlights:

• dart pub publish --dry-run
• GitHub CI on PR #237 and post-merge main
• LiteRT-LM Smoke on post-merge main
• ./tool/docs/validate_links.sh
• Tag-triggered Publish to pub.dev workflow
• Tag-triggered Docs Version Cut workflow

v0.8.5

0.8.5

Patch release for the mtmd split-library fallback ABI used by multimodal llama.cpp loads.

• Fixed the split-library mtmd fallback ABI for image and byte-buffer multimodal inputs so Windows mtmd.dll and other split mtmd native bundles use the same bitmap helper signature as the generated native binding path.
• This avoids corrupting the first mtmd bitmap-helper call for Gemma 4/MMProj style multimodal loads.
• Added native symbol regression coverage for the fallback ABI.

Validation highlights:

• dart pub publish --dry-run
• GitHub CI on PR #235
• LiteRT-LM Smoke on PR #235
• ./tool/docs/validate_links.sh

v0.8.3

0.8.3

Patch release for Windows CUDA backend discovery in native-assets builds.

• Fixed Windows CUDA backend discovery when the native asset bundle directory is
  not on the app PATH.
• llama.cpp backend modules are now loaded from their resolved bundle path in a
  way that lets colocated CUDA redistributables such as cudart64_12.dll,
  cublas64_12.dll, and cublasLt64_12.dll resolve correctly.

Validation highlights:

• dart pub publish --dry-run
• GitHub CI on PR #228 and post-merge main
• Publish to pub.dev
• Docs Version Cut
• Docs Pages

v0.8.2

• Updated the default llama.cpp native runtime pin to
  leehack/llamadart-native@b9694, regenerated matching Dart FFI bindings,
  refreshed the llamadart_llama_cpp_flutter Apple SwiftPM checksum, and
  updated the default WebGPU bridge asset pin to
  leehack/llama-web-bridge-assets@v0.1.17 (llama.cpp b9699). The WebGPU
  backend now caps unset large-model browser batches so Gemma 4 mem64 loads do
  not fall back to context-sized compute buffers.

• Added BackendGpuEnumeration.listGpuDevices({probeBackends}) (exposed via
  LlamaEngine.listGpuDevices) to enumerate GPU-class devices for offload
  selection: backend, per-backend mainGpu index, name, description, device
  id, type, and free/total memory per device. With an empty probeBackends
  only already-registered backends are inspected, so an unsupported GPU runtime
  cannot crash the process during enumeration; pass specific backends to opt
  into loading just those modules first. Web/WebGPU return an empty list.

• Added Cohere2 MoE / North Code chat-template detection and parsing so
  <|START_TEXT|> responses and <|START_ACTION|> tool-call arrays are
  handled separately from older Command-R templates.

v0.8.1

• Fixed docs references that still pointed at
  llamadart_litert_lm_flutter 0.0.1 and
  the pre-native.1 LiteRT-LM release after the 0.8.0 native pin sync moved
  LiteRT-LM Apple/runtime artifacts to v0.13.1-native.1.
• Routed native .litertlm image/audio chat parts through LiteRT-LM
  Conversation message JSON so bundles with native media processors can accept
  LlamaImageContent / LlamaAudioContent path and encoded-byte inputs without
  a separate mmproj projector.

v0.8.0

• Flutter Apple runtime packaging:
  • Split SwiftPM-linked Apple runtime packaging out of the core package into
    llamadart_llama_cpp_flutter for GGUF/llama.cpp and
    llamadart_litert_lm_flutter for .litertlm/LiteRT-LM. These companion
    packages live under packages/ in this repository and publish as separate
    pub.dev packages.
  • Removed Flutter plugin metadata from llamadart so pure Dart/native-assets
    consumers can keep using the core package without taking a Flutter SDK
    constraint.
  • Started the companion packages at 0.0.1; native pin sync bumps only the
    affected companion package patch version. Companion package publishing uses
    package-specific tags after the first manual pub.dev publish, and skips
    companion versions that already exist on pub.dev.
  • Changed unset or empty llamadart_native_runtimes to include all available
    runtime families. For Flutter iOS/macOS app builds, installed companion
    packages decide Apple SPM runtimes; for every other build,
    llamadart_native_runtimes remains the selector.
  • Updated the default llama.cpp native runtime pin to
    leehack/llamadart-native@b9587, regenerated matching Dart FFI bindings,
    and refreshed the llamadart_llama_cpp_flutter Apple SwiftPM checksum.
• MTP benchmarking diagnostics:
  • Added llama.cpp speculative decoding perf diagnostics for decode timing,
    draft/accepted token counts, draft verification timing, and acceptance
    rate so MTP benchmarks can separate backend decode cost from drafting
    overhead.
  • Extended local macOS and chat app benchmark outputs with the new
    diagnostics and added focused llama.cpp MTP smoke/benchmark tools for
    baseline-vs-MTP comparisons.
• llama.cpp MTP runtime support:
  • Added SpeculativeDecodingConfig.mtp(draftModelPath: ...) for llama.cpp
    external draft-model MTP sessions, with draft model caching and cleanup
    tied to the target model lifetime.
  • Removed the Android Vulkan MTP allow-list dart define and the model-name
    based Android Vulkan acceleration shortcut; Vulkan MTP now runs only when
    callers explicitly request Vulkan plus MTP in runtime parameters.
• Structured output:
  • Added responseFormat routing to LlamaEngine.create(...) for
    grammar-capable backends, deprecated the legacy chatTemplate(...)
jsonSchema shortcut, and made strict response-format requests fail early
    on LiteRT-LM instead of silently degrading to unconstrained generation.
• LiteRT-LM chat parity:
  • Routed eligible native .litertlm text chat through LiteRT-LM Conversation
    APIs so structured history, system messages, tool declarations, and
    template extra context reach the runtime without a Dart-rendered prompt.
    Unsupported cases still fall back to the existing Dart chat-template path.
• LiteRT-LM runtime tuning controls:
  • Added opt-in native .litertlm ModelParams for
    liteRtLmActivationDataType, liteRtLmPrefillChunkSize,
    liteRtLmParallelFileSectionLoading, and liteRtLmDispatchLibDir,
    forwarding the pinned LiteRT-LM v0.13.1 engine-settings C APIs while
    keeping defaults unchanged.
  • Extended the LiteRT-LM engine smoke tool with matching environment
    variables so real-model runs can validate load time, prefill throughput,
    decode throughput, and selected runtime settings.
  • Documented support decisions for each candidate native knob and kept
    LiteRT-LM web rejecting these native-only settings explicitly.