| Claude Fable 5.1 |
Anthropic |
Live |
1M |
$10.00 |
$50.00 |
Latest Fable - most capable Claude. Always-on thinking tuned via effort (low here). Rejects temperature. Server-side refusal fallbacks enabled. Slowest and most expensive option. |
| Claude Fable 5 |
Anthropic |
Live |
1M |
$10.00 |
$50.00 |
Most intelligent Claude. Always-on thinking. Rejects temperature parameter. Requires 30-day data retention. New tokenizer uses roughly 30 percent more tokens. |
| Claude Opus 5.5 |
Anthropic |
Live |
1M |
$4.00 |
$20.00 |
Latest Opus. Thinking cannot be disabled; effort is the control (default medium, set to low here). Rejects temperature. Server-side refusal fallbacks enabled. |
| Claude Opus 4.8 |
Anthropic |
Live |
1M |
$5.00 |
$25.00 |
Most capable Opus. WARNING: rejects temperature parameter (HTTP 400) - do not send temperature. |
| Claude Sonnet 5 |
Anthropic |
Live |
1M |
$2.00 |
$10.00 |
Current Sonnet. Faster and cheaper than Sonnet 4.6. Rejects temperature (leave NULL). Thinks by default; RequestOptions sets adaptive thinking at low effort (thinking tokens count against MaxOutputTokens). |
| Claude Sonnet 4.6 |
Anthropic |
Live |
1M |
$3.00 |
$15.00 |
Best speed/intelligence balance. Replaces claude-sonnet-4-20250514 which retires 2026-06-15. |
| Claude Haiku 4.5 |
Anthropic |
Live |
200K |
$1.00 |
$5.00 |
Fastest and cheapest Claude. 200K context. |
| GPT-5.5 |
OpenAI |
Live |
-- |
$5.00 |
$30.00 |
OpenAI flagship as of June 2026. |
| GPT-5.4 Mini |
OpenAI |
Live |
-- |
$0.75 |
$4.50 |
Mid-tier workhorse. |
| GPT-5.4 Nano |
OpenAI |
Live |
-- |
$0.20 |
$1.25 |
Cheapest current OpenAI model. |
| Grok 4.3 |
xAI (Grok) |
Configured |
1M |
$1.25 |
$2.50 |
xAI flagship - fastest and most intelligent Grok. |
| Grok 4.20 Reasoning |
xAI (Grok) |
Live |
2M |
$1.25 |
$2.50 |
Long-context reasoning variant. |
| Grok 4.1 Fast |
xAI (Grok) |
Configured |
2M |
$0.20 |
$0.50 |
Budget tier. Verify model ID with Test button before production use. |
| Gemini 3.1 Pro |
Google (Gemini) |
Configured |
2M |
$2.00 |
$12.00 |
Google flagship. PAID ONLY - returns 429 quota error on free-tier keys. Enable billing on the Google project to use. |
| Gemini 3.5 Flash |
Google (Gemini) |
Configured |
1M |
$1.50 |
$9.00 |
FREE TIER available. NOTE: intermittent 503 high-demand errors as of June 2026 (new model) - retry usually succeeds. |
| Gemini 3.1 Flash-Lite |
Google (Gemini) |
Live |
1M |
$0.25 |
$1.50 |
Cheapest Gemini. FREE TIER available. Most reliable Gemini model (default for testing). |
| Mistral Large 3 |
Mistral |
Configured |
256K |
$2.00 |
$6.00 |
Mistral flagship via -latest alias. |
| Mistral Medium 3 |
Mistral |
Configured |
128K |
$1.00 |
$3.00 |
Mid tier. |
| Mistral Small 3.1 |
Mistral |
Configured |
128K |
$0.20 |
$0.60 |
Budget tier. Free tier eligible. |
| DeepSeek V4 Pro |
DeepSeek |
Live |
128K |
$1.74 |
$3.48 |
Reasoning model. Replaces deepseek-reasoner alias (deprecated 2026-07-24). |
| DeepSeek V4 Flash |
DeepSeek |
Live |
128K |
$0.14 |
$0.28 |
Extremely cheap. Replaces deepseek-chat alias (deprecated 2026-07-24). |
| Llama 3.3 70B Versatile |
Groq |
Live |
128K |
$0.59 |
$0.79 |
Groq flagship open model. |
| GPT-OSS 120B |
Groq |
Live |
128K |
$0.15 |
$0.60 |
OpenAI open-weight model on Groq LPU. |
| Kimi K2 Instruct |
Groq |
Configured |
256K |
$1.00 |
$3.00 |
Moonshot Kimi K2 on Groq. Output price approximate - verify on groq.com/pricing. |
| Free Models Router |
OpenRouter |
Live |
128K |
Free |
Free |
FREE - auto-routes to best available free model. Ideal demo model. |
| Auto Router |
OpenRouter |
Configured |
128K |
-- |
-- |
Smart paid router - cost varies by routed model. |
| Llama 4 Maverick (Free) |
OpenRouter |
Configured |
128K |
Free |
Free |
FREE. Vision capable. Availability can change - verify with Test button. |