v0.4.1

What's Changed

• moved this post_hoc_get_action to private codebase by @optimass in #218
• Update repro o1 o3 mini by @jardinetsouffleton in #216
• Add VLLMChatModel support to chat API by @jardinetsouffleton in #220
• Add missing agent imports by @jardinetsouffleton in #223
• adding log_prob option for chat models by @TLSDC in #219
• fixing openai api typo for logprobs by @TLSDC in #224
• Switching _agents_on_benchmark to Study method for flexibility by @TLSDC in #228
• typo in hf chat models by @TLSDC in #229
• Add-claude-3.7 to repro journal by @jardinetsouffleton in #222
• fixing stale test by @TLSDC in #232
• Create a simple pure visual agent. by @recursix in #235
• Gaia bench with tape agent and multitool env by @ollmer in #214
• Generalized AgentLab Env by @recursix in #210
• adding a tag on screenshots for mouse_click coordinate actions by @TLSDC in #240
• Aj/llama4 support by @amanjaiswal73892 in #238
• fix coords tagging in agent_xray.py by @TLSDC in #242
• Allac/next agent by @recursix in #253
• Enhance_xray by @recursix in #256
• Fix count tokens by @gabrielhuang in #254
• Improve xray by @recursix in #259
• Update cost-tracking for OAI chatcompletions and response API by @amanjaiswal73892 in #260
• New experiments by @recursix in #258
• Adding a simple debug agent to manually test actions by @TLSDC in #237
• Update test_study.py by @TLSDC in #244
• Deep debug by @recursix in #262
• Multiaction Support and refactoring by @amanjaiswal73892 in #257
• OSWorld benchmark by @ollmer in #255
• Update relaunch flag to False and adjust model_args for OSWORLD_OAI by @amanjaiswal73892 in #263
• rename langchain BaseMessage by @amanjaiswal73892 in #243
• Adding new llms by @recursix in #266
• Deep debug by @recursix in #265
• Episode to html by @recursix in #267
• Fix: cache breakpoint handling in StructuredDiscussion class. by @amanjaiswal73892 in #269
• Osworld Docker Docs by @ollmer in #271
• Add a citation section by @xhluca in #274
• Xray fixes for backward (pre bgym v0.14.2) compatibility by @amanjaiswal73892 in #275
• Add Litellm API integration by @amanjaiswal73892 in #273
• Add Tutorial, GPT-5 and improve installation. by @amanjaiswal73892 in #278
• xray bugfix by @Emilianopp in #276
• Hints retrieval in tool use agent by @ollmer in #277
• Remove Langchain Dependency by @amanjaiswal73892 in #282
• Fix loading exp_dir using row_index in xray and add error handling. by @amanjaiswal73892 in #283
• Add up-to-date Azure OpenAI Agents by @patricebechard in #284
• Fix tests for Azure OpenAI models, add GPT-4.1-nano by @patricebechard in #286
• Human in the Loop Agent UI and Agent interface by @amanjaiswal73892 in #290
• Make transformer lazy import by @amanjaiswal73892 in #292
• increase timeout indefinetely in agentlab-mentor by @amanjaiswal73892 in #293
• make langchain optional in xray by @amanjaiswal73892 in #299
• Add AgentLab landing page and project pages by @imenelydiaker in #303
• Allow custom VLLM endpoint URL by @imenelydiaker in #306
• add step navigation buttons in xray. by @amanjaiswal73892 in #300
• Support for AnthropicBedrock models by @patricebechard in #307
• Hint use agent by @amanjaiswal73892 in #316
• Automated release versioning and structured PyPI/TestPyPI workflow by @amanjaiswal73892 in #308
• CUA like agent with tool use and hint support. by @amanjaiswal73892 in #318

New Contributors

• @gabrielhuang made their first contribution in #254
• @Emilianopp made their first contribution in #276
• @imenelydiaker made their first contribution in #303

Full Changelog: v0.4.0...v0.4.1