NOUMENOpen app
NOUMEN

Google · Efficient

Gemini 3.5 Flash Thinking

Added May 19, 2026

Compare models

Choose up to 3 models to compare.1 of 3 chosen

Loading models…

Use in the app

Gemini 3.5 Flash with higher thinking enabled for harder reasoning, agentic coding, tool use, multimodal analysis, and long-context workflows. Audio input costs $3.00 per million audio tokens.

  • Reasoning
  • Vision
  • Audio input
  • Video input
Input
Text, Images, Video, Audio
Output
Text
Reasoning effort
minimal · low · medium · high
Discovery tier
Efficient

Pricing

USD per 1M tokens
Input
$1.50
Output
$9.00
Cache read
$0.15
Cache write
$0.0833

Reference prices as published by the model provider. API example

Specifications

Context window
1M
Max output
65.5K
Average output (7 days)
1.7K tokens
Longer than 88% of 465 models

Side by side

Gemini 3.5 Flash Thinking compared

Specifications, prices, speed and benchmark scores next to each other. Benchmarks below show every compared model too.

Compare models

Choose up to 3 models to compare.1 of 3 chosen

Loading models…

Gemini 3.5 Flash ThinkingGoogleThis modelMuse Spark 1.1MetaRemove
ProviderGoogleMeta
Discovery tierEfficientBalanced
AddedMay 19, 2026Jul 9, 2026
Limits
Context window1.05MBest1M
Max output65.54K65.54K
Average output1.7K tokens460 tokens
Pricing, USD per 1M tokens
Input$1.50$1.25Best
Output$9.00$4.25Best
Cache read$0.15$0.15
Cache write$0.0833—
Performance
Throughput46.8 tok/s—
Time to first token8.7 s—
Request success (24h)90.1%—
Benchmark indexes
Intelligence Index32.633.7Best
Coding Index70.171.3Best
Capabilities
ReasoningYesYes
VisionYesYes
Audio inputYesYes
Video inputYesYes
Reasoning effortminimal, low, medium, highminimal, low, medium, high, xhigh

Public evaluations

Benchmarks

Independent evaluation results, with how each score ranks among the models in the Noumen catalog.

  • Gemini 3.5 Flash Thinking
  • Muse Spark 1.1

Intelligence Index

  • Gemini 3.5 Flash Thinking32.6
  • Muse Spark 1.133.7

Gemini 3.5 Flash Thinking tested at high reasoning effort

Gemini 3.5 Flash Thinking: Better than 76% of 326 scored models

3.8Catalog models by score55.1

Indexes

  • Coding Index

    Composite of the coding evaluations

    • Gemini 3.5 Flash Thinking: 70.1
    • Muse Spark 1.1: 71.3Top

Agentic work

  • AutomationBench-AA

    Workflow automation, penalized for guardrail violations

    • Gemini 3.5 Flash Thinking: 42.1%Top
    • Muse Spark 1.1: 38.8%

  • AutomationBench-AA tasks completed

    Workflows finished end to end without a guardrail violation

    • Gemini 3.5 Flash Thinking: 13.5%
    • Muse Spark 1.1: 13.7%Top

  • AA-Briefcasev1.1

    Agentic knowledge work, rated head to head

    • Gemini 3.5 Flash Thinking: 874 EloTop
    • Muse Spark 1.1: 850 Elo

  • GDPval-AAv2.1

    Economically valuable occupational tasks, rated head to head

    • Gemini 3.5 Flash Thinking: 1,202 Elo
    • Muse Spark 1.1: 1,211 EloTop

Document reasoning

  • GDP.pdf

    Professional reasoning over PDF documents, all-pass rate

    • Gemini 3.5 Flash Thinking: 19.8%Top
    • Muse Spark 1.1: 14.4%

  • AA-LCRv1.1

    Reasoning across long documents

    • Gemini 3.5 Flash Thinking: 73.3%
    • Muse Spark 1.1: 77.7%Top

  • MLCR-AA

    Long-context reasoning over medical records

    • Gemini 3.5 Flash Thinking: 10.0%Top
    • Muse Spark 1.1: 7.2%

Reasoning and knowledge

  • Humanity’s Last Exam

    Expert-level questions across academic fields

    • Gemini 3.5 Flash Thinking: 42.7%
    • Muse Spark 1.1: 46.2%Top

  • GPQA Diamond

    Graduate-level science questions

    • Gemini 3.5 Flash Thinking: 92.2%Top
    • Muse Spark 1.1: 89.8%

  • IFBench

    Following precise output instructions

    • Gemini 3.5 Flash Thinking: 76.3%
    • Muse Spark 1.1: —

Coding

  • Terminal-Benchv4.0

    Practical software engineering in a terminal

    • Gemini 3.5 Flash Thinking: 6.6%Top
    • Muse Spark 1.1: 6.1%

  • SciCode

    Scientific computing in Python

    • Gemini 3.5 Flash Thinking: 53.9%
    • Muse Spark 1.1: 58.8%Top

Legacy evaluations

  • Terminal-Bench Hard

    Earlier terminal task suite

    • Gemini 3.5 Flash Thinking: 40.9%
    • Muse Spark 1.1: —

  • τ²-Bench Telecom

    Tool use in a customer-support simulation

    • Gemini 3.5 Flash Thinking: 95.3%
    • Muse Spark 1.1: —

  • AA-LCR (unversioned)

    Long-context reasoning, earlier grading

    • Gemini 3.5 Flash Thinking: 73.3%
    • Muse Spark 1.1: 77.7%Top

Scores from Artificial Analysis, updated Oct 26, 2026. Gemini 3.5 Flash Thinking was tested at high reasoning effort; compare scores measured at the same effort.Artificial Analysis

Live routing data

Performance and reliability

Recent averages across requests routed for this model. Your own speed depends on prompt length, reasoning effort and load.

Throughput
46.8 tok/s
Faster than 34% of 250 models
Time to first token
8.7 s
Quicker than 14% of 250 models
Prompt cache hits
21.3%
About $1.21 per 1M input tokens with caching
Request success (24h)
90.2%
Unstable

Cache hits are the share of input tokens recently served from a prompt cache across all requests. They are not a forecast of your own hit rate.

Hourly request success · last 72 hoursLowest hour 60.0%
  • Healthy (98% or more)
  • Degraded (90–98%)
  • Down (under 90%)
  • No requests

Developers

Use it through the API

Noumen’s API is OpenAI-compatible: any Chat Completions client works with a base URL and a key.

Read the quickstart
curl https://www.noumen.xyz/api/v1/chat/completions \
  -H "Authorization: Bearer $NOUMEN_API_KEY" \
  -H "Content-Type: application/json" \
  -d '{
    "model": "google/gemini-3.5-flash-thinking",
    "reasoning_effort": "medium",
    "messages": [{ "role": "user", "content": "Explain this stack trace." }]
  }'
import os
from openai import OpenAI

client = OpenAI(base_url="https://www.noumen.xyz/api/v1", api_key=os.environ["NOUMEN_API_KEY"])
reply = client.chat.completions.create(
    model="google/gemini-3.5-flash-thinking",
    messages=[{"role": "user", "content": "Explain this stack trace."}],
)
print(reply.choices[0].message.content)

Google

Other Google models, starting with the same family.