diff --git a/core/config/onboarding.ts b/core/config/onboarding.ts index ed5f0019828..68a0530e158 100644 --- a/core/config/onboarding.ts +++ b/core/config/onboarding.ts @@ -3,8 +3,8 @@ import { ConfigYaml } from "@continuedev/config-yaml"; export const LOCAL_ONBOARDING_PROVIDER_TITLE = "Ollama"; export const LOCAL_ONBOARDING_FIM_MODEL = "qwen2.5-coder:1.5b-base"; export const LOCAL_ONBOARDING_FIM_TITLE = "Qwen2.5-Coder 1.5B"; -export const LOCAL_ONBOARDING_CHAT_MODEL = "llama3.1:8b"; -export const LOCAL_ONBOARDING_CHAT_TITLE = "Llama 3.1 8B"; +export const LOCAL_ONBOARDING_CHAT_MODEL = "qwen3:8b"; +export const LOCAL_ONBOARDING_CHAT_TITLE = "Qwen 3 8B"; export const LOCAL_ONBOARDING_EMBEDDINGS_MODEL = "nomic-embed-text:latest"; export const LOCAL_ONBOARDING_EMBEDDINGS_TITLE = "Nomic Embed"; diff --git a/core/llm/toolSupport.ts b/core/llm/toolSupport.ts index e38a6de8d4a..ffee2498543 100644 --- a/core/llm/toolSupport.ts +++ b/core/llm/toolSupport.ts @@ -219,6 +219,8 @@ export const PROVIDER_TOOL_SUPPORT: Record boolean> = "glm-5", "deepseek", "dolphin", + "gemma3", + "phi4", ].some((part) => modelName.toLowerCase().includes(part)) ) { return true; diff --git a/docs/customize/model-providers/top-level/ollama.mdx b/docs/customize/model-providers/top-level/ollama.mdx index 8155dba53f9..afa6e76f23d 100644 --- a/docs/customize/model-providers/top-level/ollama.mdx +++ b/docs/customize/model-providers/top-level/ollama.mdx @@ -4,13 +4,9 @@ slug: ../ollama sidebarTitle: "Ollama" --- - - **Discover Ollama models [here](https://continue.dev/lmstudio)** - +**Discover Ollama models [here](https://continue.dev/lmstudio)** - - Get started with [Ollama](https://ollama.com/download) - +Get started with [Ollama](https://ollama.com/download) ## Configuration @@ -42,10 +38,12 @@ sidebarTitle: "Ollama" } ``` + - **Check out a more advanced configuration [here](https://continue.dev/ollama/qwen3-coder-30b?view=config)** + **Check out a more advanced configuration + [here](https://continue.dev/ollama/qwen3-coder-30b?view=config)** ## How to Configure Model Capabilities in Ollama @@ -85,10 +83,12 @@ Ollama models usually have their capabilities auto-detected correctly. However, } ``` + - Many Ollama models support tool use by default. Vision models often also support image input + Many Ollama models support tool use by default. Vision models often also + support image input ## Troubleshooting @@ -99,9 +99,9 @@ Continue may set a higher default context length than other Ollama tools, causin ```yaml title="config.yaml" models: - - name: Deepseek R1 + - name: Qwen 3 8B provider: ollama - model: deepseek-r1:latest + model: qwen3:8b defaultCompletionOptions: contextLength: 2048 ``` diff --git a/docs/customize/model-roles/chat.mdx b/docs/customize/model-roles/chat.mdx index 60139c79d03..90c782c50f8 100644 --- a/docs/customize/model-roles/chat.mdx +++ b/docs/customize/model-roles/chat.mdx @@ -5,8 +5,7 @@ keywords: [chat, model, role] sidebar_position: 1 --- -import { ModelRecommendations } from '/snippets/ModelRecommendations.jsx' - +import { ModelRecommendations } from "/snippets/ModelRecommendations.jsx"; A "chat model" is an LLM that is trained to respond in a conversational format. Because they should be able to answer general questions and generate complex code, the best chat models are typically large, often 405B+ parameters. @@ -33,13 +32,13 @@ Our current top recommendations are Claude Opus 4.6 and Claude Sonnet 4 from [An version: 0.0.1 schema: v1 - models: - - name: Claude Opus 4.6 - provider: anthropic - model: claude-opus-4-6 - apiKey: - ``` - +models: - name: Claude Opus 4.6 +provider: anthropic +model: claude-opus-4-6 +apiKey: + +```` + ### Gemma from Google DeepMind @@ -47,45 +46,45 @@ Our current top recommendations are Claude Opus 4.6 and Claude Sonnet 4 from [An If you prefer to use an open-weight model, then the Gemma family of Models from Google DeepMind is a good choice. You will need to decide if you use it through a SaaS model provider, e.g. [Together](../model-providers/more/together), or self-host it, e.g. [Ollama](../model-providers/top-level/ollama). - - - - Add the [Ollama Gemma 3 27B block](https://continue.dev/ollama/gemma3-27b) from the hub - - - Add the [Together Gemma 2 27B Instruct block](https://continue.dev/togetherai/gemma-2-instruct-27b) from the hub - - - - - - - ```yaml title="config.yaml" - name: My Config - version: 0.0.1 - schema: v1 + + + + Add the [Ollama Gemma 3 27B block](https://continue.dev/ollama/gemma3-27b) from the hub + + + Add the [Together Gemma 2 27B Instruct block](https://continue.dev/togetherai/gemma-2-instruct-27b) from the hub + + + + + + + ```yaml title="config.yaml" + name: My Config + version: 0.0.1 + schema: v1 - models: - - name: "Gemma 3 27B" - provider: "ollama" - model: "gemma3:27b" - ``` - - - ```yaml title="config.yaml" - name: My Config - version: 0.0.1 - schema: v1 + models: + - name: "Gemma 3 27B" + provider: "ollama" + model: "gemma3:27b" + ``` + + + ```yaml title="config.yaml" + name: My Config + version: 0.0.1 + schema: v1 - models: - - name: "Gemma 3 27B" - provider: "together" - model: "google/gemma-2-27b-it" - apiKey: - ``` - - - + models: + - name: "Gemma 3 27B" + provider: "together" + model: "google/gemma-2-27b-it" + apiKey: + ``` + + + ### GPT-5.1 from OpenAI @@ -93,21 +92,22 @@ If you prefer to use an open-weight model, then the Gemma family of Models from If you prefer to use a model from [OpenAI](../model-providers/top-level/openai), then we recommend GPT-5.1. - - Add the [OpenAI GPT-5.1 block](https://continue.dev/openai/gpt-5.1) from the hub - - - ```yaml title="config.yaml" - name: My Config - version: 0.0.1 - schema: v1 + + Add the [OpenAI GPT-5.1 block](https://continue.dev/openai/gpt-5.1) from the hub + + +```yaml title="config.yaml" +name: My Config +version: 0.0.1 +schema: v1 + +models: + - name: GPT-5.1 + provider: openai + model: gpt-5.1 + apiKey: +```` - models: - - name: GPT-5.1 - provider: openai - model: gpt-5.1 - apiKey: - ``` @@ -125,13 +125,13 @@ If you prefer to use a model from [xAI](../model-providers/more/xAI), then we re version: 0.0.1 schema: v1 - models: - - name: Grok-4.1 - provider: xAI - model: grok-4-1-fast-non-reasoning - apiKey: - ``` - +models: - name: Grok-4.1 +provider: xAI +model: grok-4-1-fast-non-reasoning +apiKey: + +```` + ### Gemini 3.1 Pro from Google @@ -139,21 +139,22 @@ If you prefer to use a model from [xAI](../model-providers/more/xAI), then we re If you prefer to use a model from [Google](../model-providers/top-level/gemini), then we recommend Gemini 3.1 Pro. - - Add the [Gemini 3.1 Pro block](https://continue.dev/google/gemini-3.1-pro-preview) from the hub - - - ```yaml title="config.yaml" - name: My Config - version: 0.0.1 - schema: v1 + + Add the [Gemini 3.1 Pro block](https://continue.dev/google/gemini-3.1-pro-preview) from the hub + + +```yaml title="config.yaml" +name: My Config +version: 0.0.1 +schema: v1 + +models: + - name: Gemini 3.1 Pro + provider: gemini + model: gemini-3.1-pro-preview + apiKey: +```` - models: - - name: Gemini 3.1 Pro - provider: gemini - model: gemini-3.1-pro-preview - apiKey: - ``` @@ -161,20 +162,16 @@ If you prefer to use a model from [Google](../model-providers/top-level/gemini), For the best local, offline Chat experience, you will want to use a model that is large but fast enough on your machine. -### Llama 3.1 8B +### Qwen 3 8B -If your local machine can run an 8B parameter model, then we recommend running Llama 3.1 8B on your machine (e.g. using [Ollama](../model-providers/top-level/ollama) or [LM Studio](../model-providers/top-level/lmstudio)). +If your local machine can run an 8B parameter model, then we recommend running Qwen 3 8B on your machine (e.g. using [Ollama](../model-providers/top-level/ollama) or [LM Studio](../model-providers/top-level/lmstudio)). - Add the [Ollama Llama 3.1 8b block](https://continue.dev/ollama/llama3.1-8b) from the hub + Add the [Ollama Qwen 3 8B block](https://continue.dev/ollama/qwen3-8b) from the hub - {/* HUB_TODO nonexistent block */} - {/* - Add the [LM Studio Llama 3.1 8b block](https://continue.dev/explore/models) from the hub - */} @@ -186,9 +183,9 @@ If your local machine can run an 8B parameter model, then we recommend running L schema: v1 models: - - name: Llama 3.1 8B + - name: Qwen 3 8B provider: ollama - model: llama3.1:8b + model: qwen3:8b ``` @@ -198,43 +195,21 @@ If your local machine can run an 8B parameter model, then we recommend running L schema: v1 models: - - name: Llama 3.1 8B + - name: Qwen 3 8B provider: lmstudio - model: llama3.1:8b - ``` - - - ```yaml title="config.yaml" - name: My Config - version: 0.0.1 - schema: v1 - - models: - - name: Llama 3.1 8B - provider: msty - model: llama3.1:8b + model: qwen3:8b ``` + -### DeepSeek Coder 2 16B +### Qwen 3 Coder 30B -If your local machine can run a 16B parameter model, then we recommend running DeepSeek Coder 2 16B (e.g. using [Ollama](../model-providers/top-level/ollama) or [LM Studio](../model-providers/top-level/lmstudio)). +If your local machine can run a larger model, then [Qwen 3 Coder](https://ollama.com/library/qwen3-coder) is an excellent code-specialized option (e.g. using [Ollama](../model-providers/top-level/ollama) or [LM Studio](../model-providers/top-level/lmstudio)). The 30B-A3B variant uses mixture-of-experts and runs efficiently despite its size. - {/* HUB_TODO nonexistent blocks */} - {/* - - - Add the [Ollama Deepseek Coder 2 16B block](https://continue.dev/explore/models) from the hub - - - Add the [LM Studio Deepseek Coder 2 16B block](https://continue.dev/explore/models) from the hub - - - */} @@ -244,9 +219,9 @@ If your local machine can run a 16B parameter model, then we recommend running D schema: v1 models: - - name: DeepSeek Coder 2 16B + - name: Qwen 3 Coder 30B provider: ollama - model: deepseek-coder-v2:16b + model: qwen3-coder:30b-a3b ``` @@ -256,24 +231,13 @@ If your local machine can run a 16B parameter model, then we recommend running D schema: v1 models: - - name: DeepSeek Coder 2 16B + - name: Qwen 3 Coder 30B provider: lmstudio - model: deepseek-coder-v2:16b - ``` - - - ```yaml title="config.yaml" - name: My Config - version: 0.0.1 - schema: v1 - - models: - - name: DeepSeek Coder 2 16B - provider: msty - model: deepseek-coder-v2:16b + model: qwen3-coder:30b-a3b ``` + diff --git a/docs/guides/ollama-guide.mdx b/docs/guides/ollama-guide.mdx index cace7fd4c75..319a71b43cd 100644 --- a/docs/guides/ollama-guide.mdx +++ b/docs/guides/ollama-guide.mdx @@ -97,16 +97,13 @@ models: ``` - **Important**: Blocks only provide configuration - you still need to pull - the model locally. The block `ollama/deepseek-r1-32b` configures Continue - to use `model: deepseek-r1:32b`, but the actual model must be installed: - ```bash - # Check what the block expects (view on continue.dev) - # Then pull that exact model tag locally - ollama pull deepseek-r1:32b # Required for ollama/deepseek-r1-32b hub block - ``` - If the model isn't installed, Ollama will return: - `404 model "deepseek-r1:32b" not found, try pulling it first` + **Important**: Blocks only provide configuration - you still need to pull the + model locally. The block `ollama/deepseek-r1-32b` configures Continue to use + `model: deepseek-r1:32b`, but the actual model must be installed: ```bash # + Check what the block expects (view on continue.dev) # Then pull that exact + model tag locally ollama pull deepseek-r1:32b # Required for + ollama/deepseek-r1-32b hub block ``` If the model isn't installed, Ollama will + return: `404 model "deepseek-r1:32b" not found, try pulling it first` ### Method 2: Using Autodetect @@ -243,7 +240,7 @@ Choose models based on your specific needs (see [recommended models](/customize/ 2. **Chat & Reasoning**: - - `llama3.1:8b` - Latest Llama with tool support + - `qwen3:8b` - Fast and capable with tool support - `mistral:7b` - Fast and versatile - `deepseek-r1:32b` - Advanced reasoning capabilities diff --git a/extensions/vscode/config_schema.json b/extensions/vscode/config_schema.json index 6eb11282100..94e0ba7394a 100644 --- a/extensions/vscode/config_schema.json +++ b/extensions/vscode/config_schema.json @@ -248,7 +248,7 @@ "### Sagemaker\nSagemaker is AWS' machine learning platform.", "### Together\nTogether is a hosted service that provides extremely fast streaming of open-source language models. To get started with Together:\n1. Obtain an API key from [here](https://together.ai)\n2. Paste below\n3. Select a model preset\n> [Reference](https://docs.continue.dev/reference/Model%20Providers/togetherllm)", "### Novita AI\n[Novita AI](https://novita.ai?utm_source=github_continuedev&utm_medium=github_readme&utm_campaign=github_link) offers an affordable, reliable, and simple inference platform with scalable [LLM APIs](https://novita.ai/docs/model-api/reference/introduction.html), empowering developers to build AI applications. To get started with Novita AI:\n1. Obtain an API key from [here](https://novita.ai/settings/key-management?utm_source=github_continuedev&utm_medium=github_readme&utm_campaign=github_link)\n2. Paste below\n3. Select a model preset\n> [Reference](https://docs.continue.dev/reference/Model%20Providers/novita)", - "### Ollama\nTo get started with Ollama, follow these steps:\n1. Download from [ollama.ai](https://ollama.ai/) and open the application\n2. Open a terminal and run `ollama run `. Example model names are `codellama:7b-instruct` or `llama2:7b-text`. You can find the full list [here](https://ollama.ai/library).\n3. Make sure that the model name used in step 2 is the same as the one in config.json (e.g. `model=\"codellama:7b-instruct\"`)\n4. Once the model has finished downloading, you can start asking questions through Continue.\n> [Reference](https://docs.continue.dev/reference/Model%20Providers/ollama)", + "### Ollama\nTo get started with Ollama, follow these steps:\n1. Download from [ollama.ai](https://ollama.ai/) and open the application\n2. Open a terminal and run `ollama run `. Example model names are `qwen3:8b` or `deepseek-r1:14b`. You can find the full list [here](https://ollama.ai/library).\n3. Make sure that the model name used in step 2 is the same as the one in config.json (e.g. `model=\"qwen3:8b\"`)\n4. Once the model has finished downloading, you can start asking questions through Continue.\n> [Reference](https://docs.continue.dev/reference/Model%20Providers/ollama)", "### Huggingface TGI\n\n> [Reference](https://docs.continue.dev/reference/Model%20Providers/huggingfacetgi)", "### Huggingface Inference API\n\n> [Reference](https://docs.continue.dev/reference/Model%20Providers/huggingfaceinferenceapi)", "### Llama.cpp\nllama.cpp comes with a [built-in server](https://github.com/ggerganov/llama.cpp/tree/master/examples/server#llamacppexampleserver) that can be run from source. To do this:\n\n1. Clone the repository with `git clone https://github.com/ggerganov/llama.cpp`.\n2. `cd llama.cpp`\n3. Run `make` to build the server.\n4. Download the model you'd like to use and place it in the `llama.cpp/models` directory (the best place to find models is [The Bloke on HuggingFace](https://huggingface.co/TheBloke))\n5. Run the llama.cpp server with the command below (replacing with the model you downloaded):\n\n```shell\n.\\server.exe -c 4096 --host 0.0.0.0 -t 16 --mlock -m models/codellama-7b-instruct.Q8_0.gguf\n```\n\nAfter it's up and running, you can start using Continue.\n> [Reference](https://docs.continue.dev/reference/Model%20Providers/llamacpp)", diff --git a/gui/src/pages/AddNewModel/configs/models.ts b/gui/src/pages/AddNewModel/configs/models.ts index db858d6a0ac..a53eb01cdd8 100644 --- a/gui/src/pages/AddNewModel/configs/models.ts +++ b/gui/src/pages/AddNewModel/configs/models.ts @@ -380,7 +380,7 @@ export const models: { [key: string]: ModelPackage } = { }, }, ], - providerOptions: ["ollama", "lmstudio", "llama.cpp"], + providerOptions: ["lmstudio", "llama.cpp"], isOpenSource: true, }, deepseekChatApi: { @@ -421,6 +421,169 @@ export const models: { [key: string]: ModelPackage } = { providerOptions: ["deepseek"], isOpenSource: true, }, + deepseekR1Local: { + title: "DeepSeek R1", + description: + "A powerful open-source reasoning model with chain-of-thought capabilities, available in distilled sizes for local use.", + params: { + title: "DeepSeek-R1-14b", + model: "deepseek-r1:14b", + contextLength: 64_000, + }, + icon: "deepseek.png", + dimensions: [ + { + name: "Parameter Count", + description: "The number of parameters in the model", + options: { + "1.5b": { + model: "deepseek-r1:1.5b", + title: "DeepSeek-R1-1.5b", + }, + "7b": { + model: "deepseek-r1:7b", + title: "DeepSeek-R1-7b", + }, + "8b": { + model: "deepseek-r1:8b", + title: "DeepSeek-R1-8b", + }, + "14b": { + model: "deepseek-r1:14b", + title: "DeepSeek-R1-14b", + }, + "32b": { + model: "deepseek-r1:32b", + title: "DeepSeek-R1-32b", + }, + "70b": { + model: "deepseek-r1:70b", + title: "DeepSeek-R1-70b", + }, + }, + }, + ], + providerOptions: ["ollama", "lmstudio", "llama.cpp"], + isOpenSource: true, + }, + qwen3Chat: { + title: "Qwen 3", + description: + "Alibaba's latest model with hybrid thinking, strong coding and reasoning capabilities.", + params: { + title: "Qwen3-8b", + model: "qwen3:8b", + contextLength: 32_768, + }, + icon: "qwen.png", + dimensions: [ + { + name: "Parameter Count", + description: "The number of parameters in the model", + options: { + "0.6b": { + model: "qwen3:0.6b", + title: "Qwen3-0.6b", + }, + "4b": { + model: "qwen3:4b", + title: "Qwen3-4b", + }, + "8b": { + model: "qwen3:8b", + title: "Qwen3-8b", + }, + "32b": { + model: "qwen3:32b", + title: "Qwen3-32b", + }, + }, + }, + ], + providerOptions: ["ollama", "lmstudio", "llama.cpp"], + isOpenSource: true, + }, + qwen3Coder: { + title: "Qwen 3 Coder", + description: + "Alibaba's latest code-specialized model with strong multi-language programming support.", + params: { + title: "Qwen3-Coder-30b", + model: "qwen3-coder:30b-a3b", + contextLength: 32_768, + }, + icon: "qwen.png", + dimensions: [ + { + name: "Parameter Count", + description: "The number of parameters in the model", + options: { + "1.5b": { + model: "qwen3-coder:1.5b", + title: "Qwen3-Coder-1.5b", + }, + "8b": { + model: "qwen3-coder:8b", + title: "Qwen3-Coder-8b", + }, + "30b-a3b (MoE)": { + model: "qwen3-coder:30b-a3b", + title: "Qwen3-Coder-30b", + }, + }, + }, + ], + providerOptions: ["ollama", "lmstudio", "llama.cpp"], + isOpenSource: true, + }, + gemma3Chat: { + title: "Gemma 3", + description: + "Google's latest open model with strong coding and instruction-following capabilities.", + params: { + title: "Gemma3-12b", + model: "gemma3:12b", + contextLength: 32_768, + }, + dimensions: [ + { + name: "Parameter Count", + description: "The number of parameters in the model", + options: { + "1b": { + model: "gemma3:1b", + title: "Gemma3-1b", + }, + "4b": { + model: "gemma3:4b", + title: "Gemma3-4b", + }, + "12b": { + model: "gemma3:12b", + title: "Gemma3-12b", + }, + "27b": { + model: "gemma3:27b", + title: "Gemma3-27b", + }, + }, + }, + ], + providerOptions: ["ollama", "lmstudio", "llama.cpp"], + isOpenSource: true, + }, + phi4: { + title: "Phi 4", + description: + "Microsoft's compact 14B model with strong reasoning and coding performance.", + params: { + title: "Phi-4", + model: "phi4:14b", + contextLength: 16_384, + }, + providerOptions: ["ollama", "lmstudio", "llama.cpp"], + isOpenSource: true, + }, deepseekCoder2Lite: { title: "DeepSeek Coder 2 Lite", description: @@ -639,13 +802,7 @@ export const models: { [key: string]: ModelPackage } = { }, }, ], - providerOptions: [ - "ollama", - "lmstudio", - "together", - "llama.cpp", - "replicate", - ], + providerOptions: ["lmstudio", "together", "llama.cpp", "replicate"], isOpenSource: true, }, mixtralTrial: { @@ -832,7 +989,6 @@ export const models: { [key: string]: ModelPackage } = { }, ], providerOptions: [ - "ollama", "lmstudio", "together", "ovhcloud", @@ -851,7 +1007,7 @@ export const models: { [key: string]: ModelPackage } = { contextLength: 20_000, title: "Granite Code", }, - providerOptions: ["ollama", "lmstudio", "llama.cpp", "replicate"], + providerOptions: ["lmstudio", "llama.cpp", "replicate"], icon: "ibm.png", isOpenSource: true, dimensions: [ @@ -910,7 +1066,7 @@ export const models: { [key: string]: ModelPackage } = { }, }, ], - providerOptions: ["ollama", "lmstudio", "llama.cpp", "replicate"], + providerOptions: ["lmstudio", "llama.cpp", "replicate"], isOpenSource: true, }, phindCodeLlama: { @@ -922,7 +1078,7 @@ export const models: { [key: string]: ModelPackage } = { model: "phind-codellama-34b", contextLength: 4096, }, - providerOptions: ["ollama", "lmstudio", "llama.cpp", "replicate"], + providerOptions: ["lmstudio", "llama.cpp", "replicate"], isOpenSource: true, }, codestral: { diff --git a/gui/src/pages/AddNewModel/configs/providers.ts b/gui/src/pages/AddNewModel/configs/providers.ts index ee431c12fe2..6531f7e0355 100644 --- a/gui/src/pages/AddNewModel/configs/providers.ts +++ b/gui/src/pages/AddNewModel/configs/providers.ts @@ -491,7 +491,7 @@ Select the \`GPT-4o\` model below to complete your provider configuration, but n description: "One of the fastest ways to get started with local models on Mac, Linux, or Windows", longDescription: - 'To get started with Ollama, follow these steps:\n1. Download from [ollama.ai](https://ollama.ai/download) and open the application\n2. Open a terminal and run `ollama run `. Example model names are `codellama:7b-instruct` or `llama2:7b-text`. You can find the full list [here](https://ollama.ai/library).\n3. Make sure that the model name used in step 2 is the same as the one in config.json (e.g. `model="codellama:7b-instruct"`)\n4. Once the model has finished downloading, you can start asking questions through Continue.', + 'To get started with Ollama, follow these steps:\n1. Download from [ollama.ai](https://ollama.ai/download) and open the application\n2. Open a terminal and run `ollama run `. Example model names are `qwen3:8b` or `deepseek-r1:14b`. You can find the full list [here](https://ollama.ai/library).\n3. Make sure that the model name used in step 2 is the same as the one in config.json (e.g. `model="qwen3:8b"`)\n4. Once the model has finished downloading, you can start asking questions through Continue.', icon: "ollama.png", tags: [ModelProviderTags.Local, ModelProviderTags.OpenSource], packages: [ diff --git a/packages/llm-info/src/providers/os.ts b/packages/llm-info/src/providers/os.ts index c0428ad1a02..744cd1073d8 100644 --- a/packages/llm-info/src/providers/os.ts +++ b/packages/llm-info/src/providers/os.ts @@ -2,8 +2,23 @@ import { LlmInfo } from "../types.js"; export const OsLlms: Omit[] = [ { - model: "starcoder2:3b", - displayName: "StarCoder 2 3B", - contextLength: 8192, + model: "qwen3:8b", + displayName: "Qwen 3 8B", + contextLength: 32_768, + }, + { + model: "qwen3-coder:30b-a3b", + displayName: "Qwen 3 Coder 30B", + contextLength: 32_768, + }, + { + model: "deepseek-r1:14b", + displayName: "DeepSeek R1 14B", + contextLength: 64_000, + }, + { + model: "gemma3:12b", + displayName: "Gemma 3 12B", + contextLength: 32_768, }, ];