Multi-output, deepseek, new models

What's Changed

• Add Gemini 2.0 Pro model, and use 2.0 flash as a default by @ahyatt in #152
• Rename llm-make-tool-function to llm-make-tool by @danielkrizian in #153
• Fix README example of tool use by @ahyatt in #156
• Add o3 mini model by @ahyatt in #157
• llm.el (llm-chat-streaming-to-point): Add text processor callback by @ultronozm in #151
• Remove function-calls, add tool-uses to Claude capabilities by @ahyatt in #159
• Add the ability to return multiple outputs via a plist in llm calls by @ahyatt in #160
• Add Claude 3.7 Sonnet model, set to default by @ahyatt in #161
• Fix Claude streaming tool use by @ultronozm in #162
• Add deepseek, with support for a separate reasoning stream by @ahyatt in #163
• Correctly standardize capabilities name for tool use, add streaming tool use capability by @ahyatt in #164
• Add llm-models function, to list available models for a service by @ahyatt in #165
• Set keep_alive for ollama correctly, when non-standard-param set by @ahyatt in #166
• Fix instances of plist-put not in a setq by @ahyatt in #167

New Contributors

• @danielkrizian made their first contribution in #153

Full Changelog: 0.23.0...0.24.0