v0.41.0

What's Changed

• Helper attribute LiteLLMModel.provider by @jamesbraza in #364
• Renaming LiteLLMModel.router method to get_router by @jamesbraza in #366
• Fixing litellm.BadRequestError for unregistered provider by @jamesbraza in #368
• add cached token tracking and update cost calculation by @jabra in #369

Full Changelog: v0.40.1...v0.41.0

v0.40.1

What's Changed

• Added GPT 4.1 to CommonLLMNames by @jamesbraza in #359
• Corrects Aviary paper links by @lukestanley in #361
• Removed claude sonnet 3.5 by @maykcaldas in #360
• Fixing RolloutManager not calling close on envs by @jamesbraza in #363
• chore(deps): lock file maintenance by @renovate[bot] in #365
• Handle LLM request refusal by @maykcaldas in #362

New Contributors

• @lukestanley made their first contribution in #361

Full Changelog: v0.40.0...v0.40.1

v0.40.0

What's Changed

• chore(deps): lock file maintenance by @renovate[bot] in #353
• Allowing latest litellm, fixing macOS resolution by @jamesbraza in #354
• Pinning lower litellm version for OpenAI reasoning effort support by @jamesbraza in #357
• chore(deps): update actions/download-artifact action to v6 by @renovate[bot] in #358

Full Changelog: v0.39.0...v0.40.0

v0.39.0

What's Changed

• Testing image costs are reflected in LLMResult.prompt_count by @jamesbraza in #341
• Fixed flaky test_call_w_figure via better prompting by @jamesbraza in #347
• chore(deps): update actions/download-artifact action to v5 by @renovate[bot] in #343
• chore(deps): update actions/setup-python action to v6 by @renovate[bot] in #344
• chore(deps): update actions/checkout action to v5 by @renovate[bot] in #342
• chore(deps): update astral-sh/setup-uv action to v7 by @renovate[bot] in #345
• chore(deps): lock file maintenance by @renovate[bot] in #346
• Modernized CI tooling by @jamesbraza in #348
• Converting bad Message from LLM into MalformedMessageError by @jamesbraza in #350
• Basic multimodal embedding support by @jamesbraza in #349
• Pulling in BAIPP's fix that ignores UV_PYTHON by @jamesbraza in #351
• Rollout timing metadata by @sidnarayanan in #352

Full Changelog: v0.38.0...v0.39.0

v0.38.0

What's Changed

• Migrating httpx.AsyncClient to httpx_aiohttp.HttpxAiohttpClient by @jamesbraza in #339
• Cost callbacks fire only on chunks with usage by @jabra in in #340

Full Changelog: v0.37.0...v0.38.0

v0.37.0

What's Changed

• Add support for a messages sliding window by @albertbou92 in #336
• Sliding window support fixes by @albertbou92 in #337
• Reduced the react text and duplicated period by @maykcaldas in #338

Full Changelog: v0.36.0...v0.37.0

v0.36.0

What's Changed

• Make the callbacks in cost_tracker async by @AndyJCai in #335

Full Changelog: v0.35.0...v0.36.0

v0.35.0

What's Changed

• Expose LLMResponse type by @AndyJCai in #334
• pin torch 2.6 by @albertbou92 in #332

Full Changelog: v0.34.0...v0.35.0

v0.34.0

What's Changed

• Add callbacks to CostTracker to enable remote cost tracking calls to be made from fh-client by @AndyJCai in #333

New Contributors

• @AndyJCai made their first contribution in #333

Full Changelog: v0.33.1...v0.34.0

v0.33.1

What's Changed

• Handle top logprobs in litellm call by @albertbou92 in #328

Full Changelog: v0.33.0...v0.33.1