What are Gemini 4 Hydrogen's benchmark scores?

As of 13 Nov 2026, Gemini 4 Hydrogen (high effort) scores 63.4 ±1.9 on the BenchLeader Index, top 50, strongest in reasoning (70) and weakest in agents & tools (62). That rests on independent benchmark results only, and every score on this page links to its publisher.

How does Gemini 4 Hydrogen rank?

Gemini 4 Hydrogen (high effort) is #39 of 606 ranked configurations on 13 Nov 2026 (top 50). 120 other configurations score within its margin, so the exact position matters less than the band. The leader is Claude Fable 5.2 (max effort) at 72.7.

Is Gemini 4 Hydrogen better than GPT-5.4?

As of 13 Nov 2026, Gemini 4 Hydrogen (high) and GPT-5.4 (xhigh) are neck and neck on quality (63.4 vs 63.1 on the BenchLeader Index, inside the ±3.3 margin of error), and Gemini 4 Hydrogen (high) costs 1.9× less, which makes it the better buy for most workloads. The head-to-head page has the benchmark-by-benchmark detail.

When was Gemini 4 Hydrogen released?

Google released Gemini 4 Hydrogen on 12 Nov 2026. BenchLeader has results for it at 5 reasoning-effort settings.

GoogleReasoning modelNewReleased 12 Nov 2026Compare with another model

Gemini 4 Hydrogen

Launch tracker · day 2

Gemini 4 Hydrogen at launch: the first independent results

As of 13 Nov 2026, Gemini 4 Hydrogen (Google, released 12 Nov 2026) has independent results from Artificial Analysis (AA Intelligence Index v5.0 50.8), LMArena (Arena Text 1490), Vals AI (Vals Index 68.5), LiveBench (LiveBench 2.0 77.0%), Design Arena (Design Arena Website 1316) and Terminal-Bench (Terminal-Bench 56.4%). It ranks #15 of 329 models on the BenchLeader Index, at 63.4. It is listed at $1.50 in / $7.50 out per 1M tokens, with a 1M-token context window. Not yet measured by Epoch AI, ARC Prize, Scale AI and SWE-bench.

Released: 12 Nov 2026 (day 2)
Provider: Google
Price: $1.50 in / $7.50 out per 1M tokens
Context window: 1M tokens
Open weights: No
BenchLeader Index: 63.4 (#15 of 329 models)

Independent board	Results so far
Artificial Analysis	AA Intelligence Index v5.0 50.8 · AA-Omniscience v1.1 43.7 · HLE-Diamond (AA) 54.9% · more below
LMArena	Arena Text 1490 · Arena Hard Prompts 1514 · Arena Coding 1518 · more below
Vals AI	Vals Index 68.5 · Excel Modeling Benchmark v1.1 75.8% · Finance Agent v2.1 65.1% · more below
LiveBench	LiveBench 2.0 77.0% · LiveBench 2.0 Reasoning 88.6% · LiveBench 2.0 Coding 78.5% · more below
Design Arena	Design Arena Website 1316
Terminal-Bench	Terminal-Bench 56.4%

Boards usually post results within days of a launch; this page is rebuilt every six hours.

Launch comparisons

Scores from independent publishers only; benchmark figures a provider publishes about its own model are not shown.

Reasoning effort

Quality#15 of 337

63.4Index

Top 10%

+13.5 on the median

Price#186 of 253

$3.00per 1M tokens

Dearer than most

3.5× the median

Speed#21 of 215

180tokens/s

Top 10%

2.3× the median

1 cheaper ranked model scores higher: GPT-6.1 Terra (64.1, $2.00).

As of 13 Nov 2026, Gemini 4 Hydrogen by Google scores 63.4 on the BenchLeader Index (top 50 of 606 ranked configurations), with its best results in reasoning (70) and weakest in agents & tools (62). API pricing: $1.50 per 1M input tokens and $7.50 per 1M output tokens ($3.00 blended at 3:1), list price on 13 Nov 2026. It streams about 180 tokens/s (Artificial Analysis). Released 12 Nov 2026.

Key facts (as of 13 Nov 2026)

Gemini 4 Hydrogen API pricing: $1.50 per 1M input tokens and $7.50 per 1M output tokens ($3.00 blended at 3:1) (list price, 13 Nov 2026).
Gemini 4 Hydrogen streams about 180 tokens/s (Artificial Analysis, 13 Nov 2026).
Gemini 4 Hydrogen was released on 12 Nov 2026 (Google).

In detail. Gemini 4 Hydrogen is a Google proprietary reasoning model, released 12 Nov 2026. Its best configuration (high reasoning effort) ranks #39 of 606 on the BenchLeader Index at 63.4 ±1.9, in the upper half. The ± is the point: 120 other configurations score within that range, so they and this one cannot be told apart on quality alone — price and speed are what separate them. It scores highest in reasoning (70) and lowest in agents & tools (62). At $3.00 per million tokens blended it is pricier than most ranked models. Output speed of 180 tokens per second puts it in the fastest quarter, with a first answer in 16.4 s. It has been measured at 5 reasoning-effort settings; this summary describes the best-scoring one, and the tabs above switch between them. Last measured 12 Nov 2026.

Blended price: $3.00/M; $1.50 in · $7.50 out; via Artificial Analysis
Output speed: 180 tok/s; measured by Artificial Analysis
First answer: 16 s; first token 2.42 s
Context: 1.0M; via models.dev
Full answer: 63 s; median, reasoning included
Cost per task: $1.87; 8 min a task, AA Index
Tokens per task: 85k; 56% of it reasoning
Released: 12 Nov 2026

How it scores by categoryDashed line = average model (50). One step of 15 = one standard deviation.

Overall index63
Reasoning70
Coding69
Agents & tools62
Maths68
Knowledge69
Human preference68

Embed this rank

More badge styles

[![Gemini 4 Hydrogen on BenchLeader: #15 overall](https://www.benchleader.com/badge/gemini-4-hydrogen.svg?style=compact)](https://www.benchleader.com/models/gemini-4-hydrogen)

Paste it at the top of README.md, beside your other badges. The badge redraws with every data update and links back to this page.

Reasoning-effort configurations

The same model behaves differently depending on how much it is allowed to think. Each row is one setting, scored only on the benchmarks that were run at that setting. “Not stated” collects results from publishers that did not say which setting they used; for a reasoning model that is usually its thinking mode, but we do not assume it. Pick a setting here or at the top of the page to see its price, speed and category scores.

Effort	Index	Rank	Speed	First answer	Chat reply cost	Categories
low	–	–	177 tok/s	5.57 s	$0.0029	Coding 38 · Knowledge 64 · Reasoning 57
medium	–	–	176 tok/s	12 s	$0.0029	Coding 62 · Knowledge 68 · Reasoning 67
highbest	63.4	#39	180 tok/s	16 s	$0.0029	Agents & tools 62 · Coding 69 · Human preference 68 · Knowledge 69 · Maths 68 · Reasoning 70
xhigh	–	–	185 tok/s	43 s	$0.0029	Coding 70 · Knowledge 70 · Reasoning 73
not stated	–	–	122 tok/s	2.42 s	$0.0029	Composite 63

Benchmark results

See all LLM benchmarks. One column per reasoning effort. Rank is among every configuration of every model on that benchmark. Hover a score for the run it came from.

Reasoning

Benchmark	low	medium	high	xhigh	not stated	Source
Arena Hard Prompts	–	–	1514#15	–	–	Arena (LMArena)
LiveBench 2.0 Reasoningnot in indexeffort not stated	–	–	–	–	88.6%#18	LiveBench
HLE-Diamond (AA)	32.2%#49	47.7%#28	54.9%#21	58.8%#15	–	Artificial Analysis
CritPt v2 hard subset (AA)	48.0%#49	61.0%#26	61.5%#23	66.0%#18	–	Artificial Analysis

Coding

Benchmark	low	medium	high	xhigh	not stated	Source
Arena Coding	–	–	1518#43	–	–	Arena (LMArena)
Arena WebDev	–	–	1704#10	–	–	Arena (LMArena)
LiveBench 2.0 Codingnot in indexeffort not stated	–	–	–	–	78.5%#13	LiveBench
Vibe Code Bench v1.1not in index	–	–	–	87.8%#13	–	Vals AI
Terminal-Bench 4.0 (AA)not in index	6.6%#128	29.3%#68	54.0%#23	53.5%#26	–	Artificial Analysis
Design Arena Websitenot in indexeffort not stated	–	–	–	–	1316#14	Design Arena
FrontierCode 2.0 (AA)	9.3%#98	31.5%#34	36.3%#23	39.0%#15	–	Artificial Analysis

Agents & tools

Benchmark	low	medium	high	xhigh	not stated	Source
Terminal-Bench	–	–	56.4%#27	–	–	Terminal-Bench
LiveBench Long Agenticnot in indexeffort not stated	–	–	–	–	61.8%#8	LiveBench
LiveBench Browsernot in indexeffort not stated	–	–	–	–	77.1%#12	LiveBench
Finance Agent v2.1not in index	–	–	–	65.1%#12	–	Vals AI
Harvey's Legal Agent Benchmark v1.1not in index	–	–	–	11.3%#21	–	Vals AI
Legal Research Benchnot in index	–	–	–	53.9%#6	–	Vals AI
Tax Agent Benchnot in index	–	–	–	52.3%#1	–	Vals AI
Terminal-Bench 4.0 (Vals)not in index	–	–	–	54.0%#11	–	Vals AI
AutomationBenchnot in index	60.5%#51	72.0%#11	74.4%#7	76.1%#5	–	Artificial Analysis
GDP.pdf (AA)not in index	28.2%#22	32.8%#1	30.0%#13	27.6%#26	–	Artificial Analysis
AA-Briefcase v1.1not in index	947#121	1296#79	1497#45	1567#29	–	Artificial Analysis
Terminal-Bench-Science 0.1 (AA)not in index	2.4%#98	15.2%#54	41.4%#28	49.5%#21	–	Artificial Analysis

Maths

Benchmark	high	Source
Arena Maths	1495#21	Arena (LMArena)

Knowledge

Benchmark	low	medium	high	xhigh	not stated	Source
LiveBench 2.0 Data Analysisnot in indexeffort not stated	–	–	–	–	83.2%#3	LiveBench
LiveBench Researchnot in indexeffort not stated	–	–	–	–	86.8%#11	LiveBench
AA-Omniscience v1.1	36.1#34	41.9#21	43.7#14	46.3#9	–	Artificial Analysis
Excel Modeling Benchmark v1.1not in index	–	–	–	75.8%#6	–	Vals AI
AA-Omniscience v1.1: accuracynot in index	46.2%#48	48.3%#40	48.2%#43	50.5%#32	–	Artificial Analysis
AA-Omniscience v1.1: non-hallucinationnot in index	81.2%#14	87.5%#6	91.3%#4	91.5%#3	–	Artificial Analysis

Instruction following

Benchmark	high	Source
Arena Instruction Followingnot in index	1503#6	Arena (LMArena)

Human preference

Benchmark	high	Source
Arena Text	1490#15	Arena (LMArena)
Arena Creative Writingnot in index	1492#6	Arena (LMArena)
Arena Multi-turnnot in index	1476#50	Arena (LMArena)
Arena Longer Queriesnot in index	1514#8	Arena (LMArena)

Multimodal

Benchmark	not stated	Source
LiveBench Designnot in index	77.3%#11	LiveBench

Long context

Benchmark	xhigh	Source
MLCRnot in index	21.1%#47	Artificial Analysis

Composite

Benchmark	low	medium	high	xhigh	not stated	Source
LiveBench 2.0effort not stated	–	–	–	–	77.0%#9	LiveBench
AA Intelligence Index v5.0not in index	26.9#76	41.2#41	50.8#21	53.9#15	–	Artificial Analysis
Vals Indexnot in index	–	–	–	68.5#6	–	Vals AI
Vals Finance Indexnot in index	–	–	–	66.1%#2	–	Vals AI

API pricing: where to run it

Every provider serving this model through OpenRouter, with throughput and first-token latency measured on live traffic over the last 30 minutes and each provider’s own price. Purple marks the best in each column.

Provider	Speed	First token	Input $/M	Output $/M	Context	Quantisation
Google AI Studio	173 tok/s	1.58 s	$0.750	$3.75	1.0M	–
Google AI Studio	160 tok/s	2.26 s	$2.70	$13.50	1.0M	–
Google Vertex	149 tok/s	2.12 s	$1.65	$8.25	1.0M	–
Google Vertex	145 tok/s	1.40 s	$2.70	$13.50	1.0M	–
Google Vertex	137 tok/s	1.64 s	$1.50	$7.50	1.0M	–
Google AI Studio	128 tok/s	2.01 s	$1.50	$7.50	1.0M	–
Google Vertex	127 tok/s	1.38 s	$1.65	$8.25	1.0M	–

What a task costs

Estimates from list price, output speed and time to first answer for the best configuration. “With caching” assumes three-quarters of the input is served from the prompt cache at $0.075 per 1M. Reasoning tokens are not modelled.

Workload	Tokens in / out	Cost	With caching	Time
Chat reply	400 / 300	$0.0029	$0.0024	18.1 s
Summarize a 30-page report	12,000 / 600	$0.022	$0.0097	19.7 s
Code edit	6,000 / 1,500	$0.020	$0.014	24.7 s
Agentic coding session	60,000 / 4,000	$0.120	$0.056	38.6 s
Structured extraction	2,000 / 200	$0.0045	$0.0024	17.5 s

New this month

Compare Gemini 4 Hydrogen with…

Embed or cite

Show where Gemini 4 Hydrogen ranks on a model card or README. The badge redraws itself with every daily update, carries the data date and links back to this page.

[![Gemini 4 Hydrogen on BenchLeader: #15 overall](https://www.benchleader.com/badge/gemini-4-hydrogen.svg)](https://www.benchleader.com/models/gemini-4-hydrogen)

Cite this

The ranking changes daily, so a reference carries the data date. APA style:

BenchLeader. (2026, November 13). Gemini 4 Hydrogen on the BenchLeader Index (method v1.4) [Data set]. Retrieved November 13, 2026, from https://www.benchleader.com/models/gemini-4-hydrogen

BibTeX:

@misc{benchleader_gemini_4_hydrogen_20261113,
  author       = {{BenchLeader}},
  title        = {{Gemini 4 Hydrogen on the BenchLeader Index (method v1.4)}},
  year         = {2026},
  month        = nov,
  howpublished = {\url{https://www.benchleader.com/models/gemini-4-hydrogen}},
  note         = {Data as of 2026-11-13}
}

Every badge style, the embeddable top-10 widget and how to cite the Index

Questions people ask

What are Gemini 4 Hydrogen's benchmark scores?: As of 13 Nov 2026, Gemini 4 Hydrogen (high effort) scores 63.4 ±1.9 on the BenchLeader Index, top 50, strongest in reasoning (70) and weakest in agents & tools (62). That rests on independent benchmark results only, and every score on this page links to its publisher.
How does Gemini 4 Hydrogen rank?: Gemini 4 Hydrogen (high effort) is #39 of 606 ranked configurations on 13 Nov 2026 (top 50). 120 other configurations score within its margin, so the exact position matters less than the band. The leader is Claude Fable 5.2 (max effort) at 72.7.
How much does Gemini 4 Hydrogen cost per 1M tokens?: As of 13 Nov 2026, Gemini 4 Hydrogen costs $1.50 per 1M input tokens and $7.50 per 1M output tokens, $3.00 blended (three input tokens per output token).
Is Gemini 4 Hydrogen better than GPT-5.4?: As of 13 Nov 2026, Gemini 4 Hydrogen (high) and GPT-5.4 (xhigh) are neck and neck on quality (63.4 vs 63.1 on the BenchLeader Index, inside the ±3.3 margin of error), and Gemini 4 Hydrogen (high) costs 1.9× less, which makes it the better buy for most workloads. The head-to-head page has the benchmark-by-benchmark detail.
When was Gemini 4 Hydrogen released?: Google released Gemini 4 Hydrogen on 12 Nov 2026. BenchLeader has results for it at 5 reasoning-effort settings.

Data as of 13 Nov 2026. Compare these configurations.

Cite as: BenchLeader, “Gemini 4 Hydrogen: benchmarks, pricing, speed and rank”, data as of 13 Nov 2026, https://www.benchleader.com/models/gemini-4-hydrogen.