Models & pricing
USD per million tokens · Standard text input and output · Snapshot: 2026-11-12
Reference retail rates are in USD per million tokens. Contact us for volume and long-term pricing. Final rates follow your agreed plan and effective customer rate card. Confirm availability before purchase.
| Model / Model ID | Input / 1M | Output / 1M |
|---|---|---|
Claude Opus 5.5claude-opus-5-5 | $2 | $10 |
Claude Opus 4.8claude-opus-4-8 | $2 | $10 |
GPT-6 Astragpt-6-astra | $5 | $25 |
Gemini 4 Argongemini-4-argon | $1.2 | $6 |
Grok 4.7 Buildgrok-4.7 | $1.4 | $4.2 |
Claude Fable 5.1claude-fable-5-1 | $6.6 | $33 |
Claude Sonnet 5.5claude-sonnet-5-5 | $1 | $5 |
GPT-6.1 Solgpt-6.1-sol | $0.8 | $4 |
GPT-6 Solgpt-6-sol | $1 | $5 |
GPT-5.6 Solgpt-5.6-sol | $2.4 | $12 |
GPT-5.6 Terragpt-5.6-terra | $1.2 | $7.2 |
Gemini 3.8 Flashgemini-3.8-flash | $0.6 | $3 |
Gemini 3.5 Flash-Litegemini-3.5-flash-lite | $0.24 | $2 |
Grok 4.3grok-4.3 | $1 | $2 |
Grok Build 0.1grok-build-0.1 | $0.6 | $1.2 |
Grok 4.5grok-4.5 | $1.4 | $4.2 |
Grok 4.6grok-4.6 | $1.4 | $4.2 |
Grok 4.20 Non-Reasoninggrok-4.20-0309-non-reasoning | $1 | $2 |
Grok 4.20 Reasoninggrok-4.20-0309-reasoning | $1 | $2 |
Grok 4.20 Multi-Agent Betagrok-4.20-multi-agent-0309 | $1 | $2 |
Cache and additional billing
Cache reads, writes, long context and tools may have separate charges. Confirm the full rate card before purchase. Your agreed plan and effective rate card govern actual billing.
Claude Opus 4.8 reference rates per million tokens: cache read $0.10; 5-minute write $2.50; 1-hour write $4. Discounts are confirmed before purchase. Count token categories separately.
Credit & top-ups
Credit is deducted as you use model APIs. Contact Run7 to confirm a top-up amount and add credit to your existing key. There is no automatic renewal. Rate changes apply to subsequent calls at the effective rate; completed charges remain unchanged.
What is not included in this table?
Cache reads and writes, long context, web search, tools, images, audio and video may have separate charges. Features vary by model. Confirm the complete rate card and tool compatibility before purchasing; this table alone is not a complete cost estimate.
Long context & special billing
For the GPT models listed, the reference notes specify 2× input and cache rates and 1.5× output rates above 272,000 tokens per request. Grok models have 200,000-token tiers and model-specific long-context rates. Confirm thresholds, rates and access in the effective rate card. Some base prices include limited-time offers. This snapshot does not update automatically.
Use the customer portal delivered by private email with your own key to check balance, effective model rates and per-request charges. Run7 does not maintain a separate usage ledger or publish credential-bearing links.
Purchase & top up →Estimate standard text costs
Estimates uncached input and output at reference retail rates. Excludes cache, long context and other charges. Use agreed rates for negotiated plans.