--- name: thalam-models description: Pick, route, or call a THALAM model whenever the user asks for chat, coding, reasoning, long-context document work, vision, OCR, image generation, video generation, text-to-speech, transcription, embeddings, or reranking — or when the user explicitly mentions THALAM. Covers Claude, GPT, Gemini, DeepSeek, Qwen, Kimi, Llama, Mistral, FLUX, Imagen, Sora, Veo, Kling, ElevenLabs, Whisper, and 100+ other models exposed through a single OpenAI-compatible endpoint. homepage: https://thalam.ai/docs/integrations/openclaw --- # THALAM Models for OpenClaw This skill connects OpenClaw to [THALAM](https://thalam.ai), a unified API gateway covering 100+ models from Anthropic, OpenAI, Google, Meta, Alibaba (Qwen), DeepSeek, Mistral, MiniMax, Moonshot, Black Forest Labs, ElevenLabs, and more — through a single OpenAI-compatible endpoint. One API key. One bill. No lock-in. ## Setup **1. Get a THALAM API key** at https://thalam.ai (sign-up takes under a minute, no credit card). **2. Add THALAM as a model provider.** Open your config (default `~/.openclaw/openclaw.json` on Mac/Linux, `C:\Users\\.openclaw\openclaw.json` on Windows — run `openclaw onboard` first if it doesn't exist) and merge the snippet below: ```json { "models": { "mode": "merge", "providers": { "thalam": { "baseUrl": "https://api.thalam.ai/v1", "apiKey": "tl-live-...", "api": "openai-completions", "models": [ { "id": "anthropic/claude-sonnet-4.6", "name": "Claude Sonnet 4.6 (THALAM)", "reasoning": false, "input": ["text", "image"], "cost": { "input": 0.003, "output": 0.015, "cacheRead": 0, "cacheWrite": 0 }, "contextWindow": 200000, "maxTokens": 8192 }, { "id": "deepseek/deepseek-v3.2", "name": "DeepSeek V3.2 (THALAM)", "reasoning": false, "input": ["text"], "cost": { "input": 0.00027, "output": 0.0011, "cacheRead": 0, "cacheWrite": 0 }, "contextWindow": 128000, "maxTokens": 8192 } ] } } }, "agents": { "defaults": { "model": { "primary": "thalam/anthropic/claude-sonnet-4.6" }, "models": { "thalam/anthropic/claude-sonnet-4.6": {}, "thalam/deepseek/deepseek-v3.2": {} } } } } ``` **3. Validate and restart.** Run `openclaw config validate` to catch JSON errors, then `openclaw gateway restart` so the new provider takes effect. Reference any THALAM model in tasks using the prefixed id format `thalam//` (e.g., `thalam/deepseek/deepseek-v3.2`). **Important schema notes:** - The provider object does **not** accept a top-level `name` field — display names live inside each `models` array entry. - Every model id you want to call must appear in **both** `models.providers.thalam.models` (full definition) **and** `agents.defaults.models` (allowlist) — OpenClaw silently refuses to route to undeclared models. - `mode: "merge"` is required so this provider config is added alongside any others (e.g. the default OpenAI plugin) instead of replacing them. ## Model selection guide When choosing a model for a task, route by workload: ### General reasoning / complex tasks - `thalam/anthropic/claude-opus-4.7` — frontier-tier reasoning, long-form writing, agent orchestration. Use when correctness matters more than cost. - `thalam/openai/gpt-5.4` — Western flagship, strong tool use, broad multilingual including Arabic. 400K context. - `thalam/google/gemini-3.1-pro` — 2M-token context, native multimodal. Use when input is huge or includes images/audio/video. ### Cost-sensitive / high-volume - `thalam/qwen/qwen3-235b-a22b-instruct-2507` — best price/quality ratio in catalog at $0.09/1M input. Strong Arabic. Use for classification, RAG, batch translation. - `thalam/openai/gpt-5-mini` — Western volume tier with 400K context. - `thalam/google/gemma-4-31b` — Google-pedigree open model, dense, cheap. ### Reasoning (visible chain-of-thought) - `thalam/deepseek/deepseek-r1-0528` — open-weight reasoning reference. Math, coding, step-by-step logic. - `thalam/openai/o4-mini` — Western reasoning workhorse, fast and cheap. - `thalam/anthropic/claude-opus-4.7-thinking` — max-effort reasoning at frontier pricing. ### Coding - `thalam/anthropic/claude-sonnet-4.6` — default coding workhorse, strong tool use, long context, prompt caching. - `thalam/qwen/qwen3-coder-480b-a35b-instruct` — flagship open-weight coder, 256K context. - `thalam/anthropic/claude-opus-4.6-coding` — top-tier coding agent for hard SWE-bench-style tasks. ### Long context (input over 200K tokens) - `thalam/meta-llama/llama-4-scout` — 10M-token context, longest in catalog. - `thalam/moonshotai/kimi-k2.6` — 262K context, strong document analysis. - `thalam/google/gemini-3.1-pro` — 2M context with native multimodal. ### Vision / multimodal input - `thalam/openai/gpt-4o-vision` — industry default for OCR, charts, document understanding. - `thalam/qwen/qwen3-vl-235b-a22b-instruct` — open vision-language with strong Arabic OCR. ### Image generation - `thalam/openai/gpt-image-1.5` — flagship for product imagery and conversational edits. - `thalam/black-forest-labs/flux-2-pro` — photorealism leader (hands, faces, lighting). - `thalam/ideogram/ideogram-3.0` — best in-image text rendering for logos and posters. - `thalam/google/nano-banana` — fastest/cheapest conversational image edits. ### Video generation - `thalam/openai/sora-2` — flagship text-to-video. - `thalam/google/veo-3.1` — 4K output with native audio generation. - `thalam/minimax/hailuo-2.3` — fastest at $0.28/gen, strong human-figure rendering. ### Speech (TTS / STT) - `thalam/elevenlabs/eleven-v3` — flagship expressive TTS with Arabic support. - `thalam/minimax/speech-2.6-hd` — 40+ languages with native pronunciation. - `thalam/openai/whisper-large-v3-turbo` — speech-to-text across 99 languages. ### Embeddings / RAG - `thalam/google/gemini-embedding-001` — top MTEB English score, multilingual. - `thalam/cohere/embed-v4` — only multimodal text+image embedding. - `thalam/baai/bge-m3` — multilingual open embedding, $0.02/1M. - `thalam/cohere/rerank-3.5` — drop-in reranker for RAG quality. ## Default routing logic If the user does not specify a model, choose by task type: 1. **Simple Q&A or chat** → `thalam/qwen/qwen3-235b-a22b-instruct-2507` (cheapest capable) 2. **Complex reasoning, planning, multi-step analysis** → `thalam/anthropic/claude-opus-4.7` 3. **Code generation or refactor** → `thalam/anthropic/claude-sonnet-4.6` 4. **Documents over 200K tokens** → `thalam/google/gemini-3.1-pro` 5. **Image input / OCR / chart reading** → `thalam/openai/gpt-4o-vision` 6. **Image generation** → `thalam/openai/gpt-image-1.5` 7. **Video generation** → `thalam/google/veo-3.1` 8. **Speech to text** → `thalam/openai/whisper-large-v3-turbo` 9. **Text to speech** → `thalam/elevenlabs/eleven-v3` 10. **Embeddings for RAG** → `thalam/google/gemini-embedding-001` Always announce the chosen model at the start of the response so the user can override with a different one. ## Examples **User:** "Summarise this 50-page PDF" **Skill picks:** `thalam/google/gemini-3.1-pro` (long context + native PDF/multimodal) **User:** "Refactor this React component for accessibility" **Skill picks:** `thalam/anthropic/claude-sonnet-4.6` (coding workhorse) **User:** "Translate these 1000 product descriptions to Arabic" **Skill picks:** `thalam/qwen/qwen3-235b-a22b-instruct-2507` (cheap + strong Arabic) **User:** "Generate a 1080p product video from this script" **Skill picks:** `thalam/google/veo-3.1` (4K with native audio) ## Pricing Pay-per-token, no minimums, no lock-in. Full pricing at https://thalam.ai/pricing. ## Catalog updates THALAM adds models regularly. The live catalog is at https://thalam.ai/models. If a model id in this skill is no longer available, fall back to the next entry in the same category. ## Support - Docs: https://thalam.ai/docs - Catalog: https://thalam.ai/models - Status: https://thalam.ai/status - Contact: support@thalam.ai