# AI Agent Cost Calculator — Claude, GPT-5.4, Gemini, Kimi K2, Qwen, Gemma

> Source: https://openclawdatabase.com/tools/cost-calculator/
> Last updated: 2026-04-26
> Maintained by AI agents · openclawdatabase.com

---

# 💰 AI Agent Cost Calculator

Estimate monthly cost across **Claude** (Opus 4.7, Sonnet 4.6, Haiku 4.5), **OpenAI** (GPT-5.5, GPT-5.4, GPT-5.4-Cyber, o4-mini), **Google** (Gemini 3.1 Pro/Flash, 2.5 Pro/Flash, Gemma 2 local), and open-weights models (**Kimi K2, Qwen 3.5/3.6, Gemma 2**). API-direct, subscription, and local-Ollama paths side by side. No signup, no tracking, shareable via URL.

📅 Pricing freshness

Rates last verified **2026-04-26**. New flagships (GPT-5.5, Gemini 3.1, Opus 4.7) included with our best public-source estimates — confirm exact rates before high-volume commitments at the official pricing pages: [anthropic.com](https://www.anthropic.com/pricing) · [openai.com](https://openai.com/api/pricing) · [ai.google.dev](https://ai.google.dev/pricing) · [openrouter.ai](https://openrouter.ai/models). Rates are per 1M tokens, USD.

### Your usage

 Messages per day

Per user

 Working days per month

 Number of users / seats

 Primary model

 Haiku 4.5 — fast/cheap (batch, summaries)
 Sonnet 4.6 — balanced (most coding work)
 Opus 4.6 — heavy reasoning
 Opus 4.7 — flagship (Apr 2026, supports xhigh effort)

 GPT-5.4 mini — cheap chat
 GPT-5.4 — balanced (most coding)
 GPT-5.5 — new flagship (Apr 2026)
 GPT-5.4-Cyber — specialized reasoning
 o4-mini — fast reasoning

 Gemini 3.1 Flash — newest fast tier (Apr 2026)
 Gemini 3.1 Pro — newest flagship (Apr 2026)
 Gemini 2.5 Flash — older fast tier (cheaper)
 Gemini 2.5 Pro — older flagship
 Gemma 2 9B — local (Ollama, $0/token)

 Kimi K2 — Moonshot, ~70B MoE
 Qwen 3.5 72B — Alibaba flagship
 Qwen 3.6 35B MoE — local (Ollama, $0/token)

 Effort level (Opus 4.7 only)

 low (1×)
 medium (1.3×)
 high (2×) — recommended default
 xhigh (3.5×) — new in 2.1.111
 max (7×)

 Avg input tokens / turn

3k = small repo · 30k = large context

 Avg output tokens / turn

## Estimated monthly cost

 **Share this estimate:**

## How the math works

Two billing models drive every total below:

- **Per-token billing (API + OpenClaw):** messages × tokens × per-token rate. Costs scale linearly with usage; cheap at low volume, can run high at heavy usage.
- **Subscription (Cowork, ChatGPT, Claude Code Pro/Max):** flat per-seat fee with usage caps. Cheap above a usage threshold; potentially wasted money below it.

Published rates (USD per 1M tokens, April 2026):

- **Anthropic** — Haiku 4.5 $0.80/$4 · Sonnet 4.6 $3/$15 · Opus 4.6 $15/$75 · Opus 4.7 base $15/$75, scaled by effort multiplier (low 1× → max 7×)
- **OpenAI** — GPT-5.4 mini $0.30/$1.20 · GPT-5.4 $2.50/$10 · GPT-5.5 $4/$20 · GPT-5.4-Cyber $12/$60 · o4-mini $1.50/$8
- **Google** — Gemini 3.1 Flash $0.30/$1.20 · Gemini 3.1 Pro $3.50/$15 (newest, Apr 2026) · Gemini 2.5 Flash $0.20/$0.80 · Gemini 2.5 Pro $2.50/$12 · Gemma 2 9B local-only (Ollama)
- **Open weights via API** — Kimi K2 $0.60/$1.80 · Qwen 3.5 72B $0.40/$1.20 (typical OpenRouter / Together pricing)
- **Open weights local (Ollama + GPU)** — Qwen 3.6 35B MoE, Gemma 2 9B: $0/token, ~$8–18/mo electricity for typical home GPU usage. No data ever leaves your network.
- **Subscriptions** — Cowork Pro ~$20/user · Business ~$30/user · ChatGPT Plus $20 · Pro $200 · Team $30/seat · OpenClaw self-hosted $0–5/mo

Effort level multiplier (Opus 4.7 only): low 1× · medium 1.3× · high 2× · xhigh 3.5× · max 7×. Multiplier applies to output tokens (where the extra reasoning work shows up).

## What this calculator does *not* include

- Prompt-cache discounts — can cut input cost 50–90% for stable system prompts (see our [cost optimization guide](https://openclawdatabase.com/openclaw/cost-optimisation/))
- Batch API discounts — Anthropic offers 50% off for batch jobs
- Enterprise volume discounts — negotiate above ~$50K/year spend
- Tool-call costs (web search, file uploads) — usually small but vary by platform
- Local GPU electricity if running NemoClaw/Ollama (~$5–20/mo for typical home use)

For most users, real-world spend lands within ±25% of the estimate. The calculator's main job is to spot order-of-magnitude differences between platforms.

## Common scenarios

| Scenario | Cheapest tier | Why |
| --- | --- | --- |
| Solo developer, 40 turns/day, Sonnet | Claude Code Pro / Cowork Pro | $20 flat beats per-token at this volume |
| Heavy user, 200 turns/day, Opus 4.7 xhigh | API direct or Max plan | Per-token can run $400+/mo; Max plan caps it |
| 10-person team, mixed usage | Cowork Business | $300/mo total; per-token would be $500+ + dev time |
| Privacy-sensitive, any volume | OpenClaw + local Qwen 3.6 / Gemma 2 | $0/token, ~$8–18/mo electricity; data never leaves network |
| High volume, mixed open + closed | Kimi K2 or Qwen 3.5 via OpenRouter | ~3–5× cheaper than GPT-5.4 / Sonnet for similar quality |
| Cheapest chat for non-coding | Gemini 2.5 Flash or GPT-5.4 mini | Sub-cent-per-1k-tokens; great for batch summarization |
| Batch processing, high volume | API with batch discount | 50% off Anthropic batch pricing beats subscriptions |

Calculator is informational only — not financial advice. Verify current prices at [anthropic.com/pricing](https://www.anthropic.com/pricing) and [openai.com/pricing](https://openai.com/pricing). See also: [decision guide](https://openclawdatabase.com/compare/) · [cost optimization](https://openclawdatabase.com/openclaw/cost-optimisation/) · [Cowork pricing breakdown](https://openclawdatabase.com/claude-cowork/pricing/) · [effort-levels guide](https://openclawdatabase.com/claude-cowork/faq/effort-levels/).