How to Use Gemini 3.5 Thinking and Fast API for Free

Gemini 3.5 is now available on Hypereal in two coding-friendly variants:

• gemini-3.5-thinking for deeper reasoning, review, and multi-step coding work
• gemini-3.5-fast for low-latency chat, test generation, and agent loops

Both models are included in the Coding Plan. The Coding Plan price is $0.90 per 1M input tokens and $5.40 per 1M output tokens — a flat 40% below the Google official price of $1.50 / $9.00.

New accounts receive free credits, so you can test the API before buying a Coding Credits pack.

Step 1: Create a free Hypereal account

Sign up at hypereal.cloud, open the dashboard, and create an API key under Manage API Keys.

Store it as an environment variable:

export HYPEREAL_API_KEY="ck_..."

Step 2: Call Gemini 3.5 Thinking

Use Thinking when you want careful reasoning over code, diffs, logs, or long task descriptions.

curl https://hypereal.cloud/v1/chat/completions \
  -H "Authorization: Bearer $HYPEREAL_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "gemini-3.5-thinking",
    "messages": [
      {
        "role": "user",
        "content": "Review this API handler and identify the highest-risk bug."
      }
    ]
  }'

Step 3: Call Gemini 3.5 Fast

Use Fast when latency matters more than maximum depth.

curl https://hypereal.cloud/v1/chat/completions \
  -H "Authorization: Bearer $HYPEREAL_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "gemini-3.5-fast",
    "messages": [
      {
        "role": "user",
        "content": "Write Jest tests for a date parser with edge cases."
      }
    ]
  }'

Step 4: Use the OpenAI SDK

Hypereal is OpenAI-compatible for chat completions:

from openai import OpenAI
import os

client = OpenAI(
    api_key=os.environ["HYPEREAL_API_KEY"],
    base_url="https://hypereal.cloud/v1",
)

resp = client.chat.completions.create(
    model="gemini-3.5-thinking",
    messages=[
        {"role": "user", "content": "Plan a safe migration for this schema change."}
    ],
)

print(resp.choices[0].message.content)

Pricing and credits

Model	Hypereal input	Hypereal output	Google official benchmark	Notes
gemini-3.5-thinking	$0.90 / 1M	$5.40 / 1M	$1.50 / $9.00	Best for reasoning
gemini-3.5-fast	$0.90 / 1M	$5.40 / 1M	$1.50 / $9.00	Best for speed

Free trial credits are enough for initial API tests. For production coding workloads, buy Coding Credits from the dashboard. Coding Credits are reserved for Claude Opus, Claude Sonnet, GPT-5.5, Gemini 3.5 Thinking, and Gemini 3.5 Fast.

Which one should you pick?

Choose Gemini 3.5 Thinking for architecture review, debugging plans, codebase reasoning, and tasks where a slower answer is acceptable.

Choose Gemini 3.5 Fast for autocomplete-style loops, quick test generation, chat, and agents that need many small calls.

Both use the same API key, the same endpoint, and the same billing system.