Release report · November 2026Gemini 4 Neon

A new high score, for less than the last one.

モデル
gemini-4-neon
供給者
Google
実行
Nov 17 – Nov 18, 2026
Benchmark
AutomationBench 1.0.6
タスク
657 per effort · 4 efforts

Gemini 4 Neon completes 56.3% of AutomationBench's strict, end-to-end business tasks at xhigh effort, the highest score on the benchmark so far. That is 2.4 points above Gemini 4 Carbon, and at high effort Neon edges past Carbon's best for 55% less per task.

Best pass rate56。3%XHigh effort · 370 of 657 tasks
vs. Gemini 4 Carbon+2.4pts53.9% at High
Cost per task$1.69XHigh effort · list-price estimate, with caching
Cheaper than Gemini 4 Carbon55%$0.88 vs. $1.94 a task, Neon at high
On the frontier4/ 4effort settings no model in this comparison beats on both cost and accuracy

Key findings

01

The highest score on the benchmark

At xhigh effort Gemini 4 Neon passes 370 of 657 tasks, 56.3%. That is 2.4 points above Gemini 4 Carbon's 53.9% and 3.9 above GPT-6.2 Astra's 52.4%, at $1.69 a task, 13% less than Carbon in this comparison.

02

The previous leaders' scores, for less

High effort reaches 54.3% for $0.88 a task, above Carbon's best for 55% less. Medium reaches 53.3% for $0.58, above GPT-6.2 Astra at max for 30% less.

03

Low effort gives up little

Going from low to xhigh adds 13% more passes, 326 to 370, while cost per task rises 320%, from $0.40 to $1.69. Every Neon setting sits on the cost and accuracy frontier.

04

Ahead in three domains, level in a fourth

At its best effort per domain, Neon leads Marketing (68.0% vs. Carbon's 64.0%), Finance (65.0% vs. 60.0%) and Sales (59.0% vs. Gemini 4 Argon's 58.1%), and ties Carbon in Operations at 68.0%, which it reaches at low effort. HR is the gap: 37.5% at best, 5.8 points behind GPT-6.2 Astra.

01

Results by effort

A shallow curve: 49.6% at low, 56.3% at xhigh.

Scoring is strict. A task passes only if every assertion about the final state of the simulated company holds. All 657 tasks stay in the denominator at every effort. The dashed line marks Gemini 4 Carbon at high effort, the strongest other result in this comparison; Neon clears it from high effort up.

Low
49。6%
326/657$0.40
Medium
53。3%
350/657$0.58
High
54。3%
357/657$0.88
XHigh
56。3%
370/657$1.69

Official pass rate · task_completed_correctly · 657 tasks per effort, all in the denominator

02

Against the field

2.4 points above Gemini 4 Carbon, the previous high.

Each bar is a model's best effort setting on AutomationBench 1.0.6. Gemini 4 Neon is shown at xhigh effort; the other results are the ones published on zapier.com/benchmarks. Claude Fable 5.2 and Gemini 4 Hydrogen appear only here, since they aren't in this run's cost comparison.

Gemini 4 NeonXHigh
56。3%
Gemini 4 CarbonHigh
53。9%
GPT-6.2 AstraMax
52。4%
Claude Fable 5.2Max
51。4%
Gemini 4 ArgonHigh
51。3%
Gemini 4 HydrogenHigh
51。1%
GPT-6.2 SolHigh
41。2%

Best observed pass rate per model · AutomationBench 1.0.6

03

Cost vs. accuracy

Four of the five frontier points are Neon's.

Each line is one model's effort settings, pass rate against cost per task. GPT-6.2 Sol at high is the cheapest result, at $0.21 a task; from $0.40 up, the cheapest way to any score in this comparison is a Neon setting. Medium, at $0.58, beats GPT-6.2 Astra at max on both cost and accuracy.

0%10%20%30%40%50%60%$0$1$2Cost per task (USD) →Gemini 4 CarbonGPT-6.2 AstraGemini 4 ArgonGPT-6.2 SolGemini 4 NeonLowMediumHighXHigh
Gemini 4 Neon Frontier Frontier point DominatedHover a point for details
View the data
モデルLowMediumHighXHigh
Gemini 4 Neon49.6%$0.4053.3%$0.5854.3%$0.8856.3%$1.69
Gemini 4 Carbon——53.9%$1.94—
GPT-6.2 Astra————
Gemini 4 Argon——51.3%$1.63—
GPT- 6 . 2ソル——41.2%$0.21—

API cost per task, list-price estimate, with caching · pass rate at each effort, low to max · Cost comparisons use the models in this run's metrics export. Claude Fable 5.2 and Gemini 4 Hydrogen are pass-rate results from the zapier.com/benchmarks leaderboard; Claude Fable 5.2's uses Anthropic's default fallbacks.

04

By domain

Marketing gains the most, and Operations doesn't move.

Each line runs across the four effort settings. Marketing climbs 13.0 points from low to xhigh, 55.0% to 68.0%. Operations scores 68.0% at low, medium and xhigh, with a dip to 65.0% at high. HR is the hardest domain at every effort and ends at 35.0%.

0%10%20%30%40%50%60%70%80%
Low → XHigh
マーケティング
55。0%68。0%
+13.0pts
オペレーションズ
68。0%68。0%
+0.0pts
金融
57。5%65。0%
+7.5pts
販売
47。9%59。0%
+11.1pts
支える
37。0%45。0%
+8.0pts
HR
34。2%35。0%
+0.8pts
Low Medium, High XHigh

Full matrix

ドメインLowMediumHighXHigh
販売47.9%56/11752.1%61/11753.8%63/11759.0%69/117
マーケティング55.0%55/10059.0%59/10064.0%64/10068.0%68/100
オペレーションズ68.0%68/10068.0%68/10065.0%65/10068.0%68/100
支える37.0%37/10041.0%41/10043.0%43/10045.0%45/100
金融57.5%69/12065.0%78/12064.2%77/12065.0%78/120
HR34.2%41/12035.8%43/12037.5%45/12035.0%42/120

Pass rate · passed / tasks per cell · darker = higher

05

費用

$0.58 a task at medium, $1.69 at xhigh.

Low has the lowest spend per passed task, $0.81, against $3.01 at xhigh. At xhigh a failed task costs $2.21 and a passed one $1.29, so misses drive the bill. Spend per pass counts the cost of every failed task too. These are estimates at Google's stated rates with an assumed cache-read rate, not billed costs.

EffortCost / taskPassed taskFailed taskSpend per pass
Low$0.40$0.44$0.37$0.81
Medium$0.58$0.60$0.55$1.08
High$0.88$0.87$0.90$1.62
XHigh$1.69$1.29$2.21$3.01

USD, list-price estimate, with caching · spend per pass = all task costs ÷ passes, failures included

06

Tokens & tools

Reasoning grows 8.8×, steps grow 64%.

Tool calls rise from 57.6 per task at low to 106.3 at xhigh, and steps from 10.4 to 17.1. Reasoning tokens grow from 6,657 to 58,544 per task and make up 82% of output at xhigh. One task, at high effort, hit the 50-step cap.

Reasoning Visible output (text + tool calls)Mean output tokens per task
Low
14.5k6.7k reasoning
Medium
21.3k12.1k reasoning
High
32.8k21.7k reasoning
XHigh
71.2k58.5k reasoning
EffortTool callsステップCalls / stepInput tokensCached shareAt step cap
Low57.610.45.5272.3k81。3%0
Medium66.111.45.8411.3k83。3%0
High82.914.35.8751.6k87。2%1
XHigh106.317.16.21310.8k88。5%0

Mean per task, all tasks included · step cap 50

07

Methodology

AutomationBench tests AI agents on end-to-end workflow execution using 47 simulated apps across six business functions: Sales, Marketing, Operations, Support, Finance and HR. Each task boots a small simulated company, hands the agent one request a real operations person would get, then grades the world it leaves behind with deterministic assertions.

  1. 01

    Boot a tiny company

    CRM records, inbox threads, spreadsheets and support cases, including traps like stale rows and near-duplicate names.

  2. 02

    The agent works alone

    The agent gets one trigger message and tools to search, read, update and send. It can't ask clarifying questions, and there's no human in the loop.

  3. 03

    Grade the end state

    Deterministic assertions check what changed. The task passes only if every one holds.

モデル
gemini-4-neon
Configuration
Gemini 4 Neon
Benchmark
AutomationBench 1.0.6
Run dates
Nov 17 – Nov 18, 2026
タスク
657 per effort setting
Efforts
Low, Medium, High, XHigh
Mode
API · 50-step cap
メトリック
task_completed_correctly (strict)

CostEstimates at Google's stated rates for Gemini 4 Neon: $3.00 input and $15.00 output per 1M tokens, with reasoning billed as output, and a $0.15 cache-read rate assumed until Google publishes one. Gemini implicit caching records no cache writes. These are not billed costs.

ComparisonsCost comparisons use the models in this run's metrics export. Claude Fable 5.2 and Gemini 4 Hydrogen are pass-rate results from the zapier.com/benchmarks leaderboard; Claude Fable 5.2's uses Anthropic's default fallbacks.