One API key on SkillBoss for all of them. Zero markup. Free $0.50 trial, no signup. The numbers below are the real public rates — what you pay through SkillBoss is exactly that, no platform fee.
Prices are per 1M tokens (input / output). Context window in tokens. Last updated April 2026.
| Model | Vendor | Context | Input $/M | Output $/M | Best for | SkillBoss ID |
|---|---|---|---|---|---|---|
| Claude 4.6 Sonnet | Anthropic | 200K | $3.00 | $15.00 | Hardest reasoning + coding | bedrock/claude-4-6-sonnet |
| Claude 4.6 Haiku | Anthropic | 200K | $1.00 | $5.00 | Fast everyday chat + tools | bedrock/claude-4-6-haiku |
| Claude 4.5 Opus | Anthropic | 200K | $5.00 | $25.00 | Longform analysis & writing | bedrock/claude-4-5-opus |
| GPT-5 | OpenAI | 272K | $1.25 | $10.00 | General intelligence | openai/gpt-5 |
| GPT-5 Mini | OpenAI | 272K | $0.25 | $2.00 | Speed + cost sweet spot | openai/gpt-5-mini |
| GPT-4.1 | OpenAI | 128K | $2.00 | $8.00 | Standard agent work | openai/gpt-4.1 |
| GPT-4.1 Nano | OpenAI | 128K | $0.10 | $0.40 | Ultra-cheap simple tasks | openai/gpt-4.1-nano |
| o3 | OpenAI | 200K | $15.00 | $60.00 | Hardest reasoning (math, science) | openai/o3 |
| Gemini 2.5 Pro | 1M | $1.25 | $5.00 | Huge context, multimodal | gemini/gemini-2.5-pro | |
| Gemini 2.5 Flash | 1M | $0.075 | $0.30 | Cheapest large-context | gemini/gemini-2.5-flash | |
| DeepSeek V3.2 | DeepSeek | 128K | $0.14 | $0.28 | Value + prompt caching | deepseek/deepseek-v3.2 |
| Llama 4 Maverick | Meta | 128K | $0.40 | $1.60 | Open-source leader | meta/llama-4-maverick |
| Grok 4 | xAI | 128K | $5.00 | $15.00 | Real-time search + edgy tone | xai/grok-4 |
| Qwen 2.5 72B | Alibaba | 128K | $0.18 | $0.54 | Multilingual (CJK strong) | qwen/qwen-2.5-72b |
| Perplexity Sonar Pro | Perplexity | 127K | $3.00 | $15.00 | Search-grounded answers | perplexity/sonar-pro |
| Mistral Large 2 | Mistral | 128K | $2.00 | $6.00 | EU / privacy-first | mistral/mistral-large-2 |
| Command R+ | Cohere | 128K | $2.50 | $10.00 | RAG + tool-use workloads | cohere/command-r-plus |
Every model in the matrix speaks the OpenAI chat-completions protocol. Just swap the model field — one API key, no per-vendor accounts.
curl https://api.heybossai.com/v1/chat/completions \
-H "Authorization: Bearer $SKILLBOSS_API_KEY" \
-H "Content-Type: application/json" \
-d '{
"model": "bedrock/claude-4-6-sonnet",
"messages": [
{ "role": "user", "content": "Refactor this function for clarity: ..." }
]
}'Swap bedrock/claude-4-6-sonnet for any SkillBoss ID in the matrix above. Same endpoint. Same key.
Gemini 2.5 Flash is the cheapest large-context model at $0.075 / $0.30 per 1M input/output tokens. GPT-4.1 Nano is cheapest for short prompts at $0.10 / $0.40. DeepSeek V3.2 ($0.14 / $0.28) is the best value for coding and reasoning with prompt caching. All three are available through one SkillBoss API key at zero markup.
Claude 4.6 Sonnet is the current state of the art for coding and agentic workflows. For budget coding, DeepSeek V3.2 is within a few points on most benchmarks at ~20x lower cost. SkillBoss lets you route between them per request — use Sonnet for hard problems and DeepSeek for bulk edits without switching API keys.
Google Gemini 2.5 Pro and Flash both ship a 1,000,000-token context window — the largest in production. GPT-5 offers 272K, Claude models offer 200K, and most open-source models cap around 128K. For codebase-scale or document-scale context, Gemini 2.5 is the default choice.
Yes. SkillBoss exposes every model in this matrix through one OpenAI-compatible endpoint at https://api.heybossai.com/v1. Just change the "model" field in the request body — no extra accounts, no extra API keys, no monthly commitments. Claude Code, Cursor, Windsurf, Cline, and any OpenAI SDK work out of the box.
Grab the $0.50 free anonymous wallet at skillboss.co (no signup, no card) to test any model in this matrix. When you like it, upgrade to a full account at skillboss.co/console — still zero markup, usage-based, and the same API key works across 100+ models.
One SkillBoss key routes to Claude, GPT, Gemini, DeepSeek, Llama, Grok and 100+ more. Zero markup.
AI Agents
Tell your agent:
set up skillboss.co/skill.mdAuto-configures base URL, auth, and model access. Works with Claude Code, Cursor, Windsurf, and any MCP-compatible agent.
Developers
1. Get API key at skillboss.co/console
2. POST to api.skillboss.co/v1/run
3. Pick a model from 600+ APIs
Pay-as-you-go. $2 free credit. No subscription required.