Gemini 4 Hydrogen — Model Card
Report an issue with this entry
Version history
| Fetched | Content type | Size | Fingerprint | Original | What changed |
|---|---|---|---|---|---|
| 2026-11-20T06:21:32Z | application/pdf | 349,786 | bf8ea4a12fcb9cea |
archived copy | The Gemini 4 Hydrogen Safety & Security Report, previously marked coming soon, is now published and linked from the card: the Safety & Security section now reads Please see the full Safety & Security Report for Gemini 4 Hydrogen here, and the frontier-safety and mitigation sections drop their (coming soon) notes. The CCL summary table is unchanged apart from reflow. |
| 2026-11-13T06:23:21Z | application/pdf | 345,732 | 54bec25be9cdc436 |
archived copy | initial version |
Provenance
Show 3 changelog entries (when, how, and why this entry changed)
-
2026-11-13T06:23:21Z
add
run 2026-11-13T06:17Z-local · agent · lead: phase_a
Official Google DeepMind model card for Gemini 4 Hydrogen, a new Gemini 4 series model launched 2026-11-12 to all Gemini API, AI Studio, Gemini app and Search users (announced in Google's 'Introducing Gemini 4 Hydrogen' post). It is a distinct release (adapted from Gemini 4 Carbon but a separately named, priced, and deployed model) with its own card. The card reports internal automated safety evaluation deltas versus Gemini 3.8 Flash (text, multilingual, image-to-text safety, tone, unjustified refusals) and a Frontier Safety Framework assessment: an MCC assessment versus Gemini 4 Argon plus targeted CBRN (biological) and cybersecurity critical-capability testing, finding the CCLs not reached but the alert thresholds reached in both domains. The full Safety & Security Report is marked 'coming soon' and should be proposed separately when published. The PDF is the full card; the deepmind.google HTML page is the web version with the same content. Publication date taken from the launch post byline (Nov 12, 2026); the card itself says 'Published: November 2026'.Criteria: document_retrievable=pass; min_publication_date=pass; publisher_on_allowlist=pass · attested about_a_specific_model_or_eval=true · attested covered_model_class=true · attested distinct_model_release=true · attested notable_release=true · attested primary_source=true
-
2026-11-20T06:21:32Z
new_version
run 2026-11-20T06:17Z-local · monitor · lead: monitor
Content fingerprint changed on scheduled re-fetch
-
2026-11-20T06:26:46Z
field_update
run 2026-11-20T06:17Z-local · agent · lead: manual
change_summary: null → "The Gemini 4 Hydrogen Safety \u0026 Security Report, previously marked coming soon, is now published and linked from the card: the Safety \u0026 Security section now reads Please see the full Safety \u0026 Security Report for Gemini 4 Hydrogen here, and the frontier-safety and mitigation sections drop their (coming soon) notes. The CCL summary table is unchanged apart from reflow."Diff v1167 to v1247 shows every 'coming soon' reference to the Safety & Security Report replaced with a live reference; the risk-domain table content is identical, only reflowed.
Extracted text
Show extracted text (from the latest fetched version; formatting is lossy — use the original for reading)
Gemini 4
Hydrogen
Model card
Gemini 4 Hydrogen — Model Card
Model Cards are intended to provide essential information on Gemini models, including
known limitations, mitigation approaches, and safety performance. Model cards may be
updated from time to time; for example, to include updated evaluations as the model is
improved or revised. See the Google DeepMind site for a comprehensive list of model
cards.
Published: November 2026
Model Information
Description
Gemini 4 Hydrogen brings cutting-edge intelligence to everyone — with speed and
efficiency for everyday coding and agentic workflows. While our frontier models, Gemini
4 Carbon and Gemini 4 Argon, are built for deep, long-horizon reasoning on your hardest
problems, Hydrogen is built for rapid execution across diverse agentic workflows.
Model dependencies
Gemini 4 Hydrogen is adapted from Gemini 4 Carbon.
Inputs
Text strings (e.g., a question, a prompt, document(s) to be summarized), images, audio,
and video files, with a combined input-output token context window of up to 1M.
Outputs
1M unified context window with up to 64K default synchronous output tokens / up to 1M
via background long decode.
Architecture
Gemini 4 Hydrogen is adapted from Gemini 4 Carbon. For more information about the
model architecture for Gemini 4 Hydrogen, see the Gemini 4 Carbon model card.
1
Model Data
Training Dataset
Gemini 4 Hydrogen is adapted from Gemini 4 Carbon. For more information about the
training dataset for Gemini 4 Hydrogen, see the Gemini 4 Carbon model card.
Training Data Processing
Gemini 4 Hydrogen is adapted from Gemini 4 Carbon. For more information about the
training data processing for Gemini 4 Hydrogen, see the Gemini 4 Carbon model card.
Implementation and Sustainability
Hardware
Gemini 4 Hydrogen is adapted from Gemini 4 Carbon. For more information about the
hardware for Gemini 4 Hydrogen and our continued commitment to operate sustainably,
see the Gemini 4 Carbon model card.
Software
Gemini 4 Hydrogen is adapted from Gemini 4 Carbon. For more information about the
software for Gemini 4 Hydrogen, see the Gemini 4 Carbon model card.
2
Distribution
Gemini 4 Hydrogen is distributed in the following channels; respective documentation
shared in line:
● Gemini app
● Gemini API
● Google Sheets
● Google AI Studio
● Gemini for business
● Google Antigravity
● Google Search
Our models are available to downstream providers via an application program interface
(API) and subject to relevant terms of use. There is no required hardware or software to
use the model. For AI Studio and Gemini API, see the Gemini API Additional Terms of
Service; for Gemini for business, see Google Cloud Platform Terms of Service. For more
information, see Gemini Model API instructions, Gemini API quickstart, and Google Cloud
documentation.
Evaluation
Approach
We evaluated Gemini 4 Hydrogen across a range of benchmarks, including agentic
coding, knowledge work, computer use, multimodal understanding, science and math,
and reasoning.
Additional benchmarks and details on approach, results and their methodologies can be
found at: deepmind.com/models/evals-methodology/gemini-4-hydrogen.
3
Results
Gemini 4 Hydrogen results as of November, 2026 are below:
4
Intended Usage and Limitations
Benefit and Intended Usage
Gemini 4 Hydrogen brings cutting-edge intelligence to everyone — with speed and
efficiency for everyday coding and agentic workflows.
It is particularly well-suited for applications that require:
● fast, disciplined execution on its own or as sub-agents orchestrated alongside
frontier models
● iterative repository coding and debugging
● open-ended agentic execution
● multimodal knowledge work
● web development
● adaptable speed and intelligence across four thinking levels
● ready-to-go writing, research, and everyday work
● connecting the dots and getting more tasks done
● everyday prototyping and iteration
Known Limitations
Gemini 4 Hydrogen may exhibit some general limitations such as hallucinations. In
addition to this, we are continually working to improve jailbreak resistance and have
further strengthened mitigations across Frontier Safety. There may also be occasional
slowness or timeout issues, and at times the model may use more tokens to maximize
performance, especially at higher thinking levels. The knowledge cutoff for Gemini 4
Hydrogen is March 2026, though in some domains the model’s knowledge may be less up
to date.
5
Acceptable Usage
Gemini 4 Hydrogen is adapted from Gemini 4 Carbon. For more information about the
acceptable usage for Gemini 4 Hydrogen, see the Gemini 4 Carbon model card.
Safety & Security
Please see the full Safety & Security Report for Gemini 4 Hydrogen here.
Evaluation Approach
For more information about the evaluation approach for Gemini 4 Hydrogen, see the
Gemini 4 Carbon model card.
Safety Policies
For more information about the safety policies for Gemini 4 Hydrogen, see the Gemini 4
Carbon model card.
Training and Development Evaluation Results
Results for some of the internal safety evaluations conducted during the development
phase are listed below. The evaluation results are for automated evaluations and not
human evaluation or red teaming. Scores are provided as an absolute percentage
increase or decrease in performance compared to the indicated model, as described
below.
Overall, Gemini 4 Hydrogen outperforms Gemini 3.8 Flash on multimodal content safety,
while text-to-text and multilingual safety remain at a similarly high safe response rate
within confidence intervals. Refusal tone is similar to Gemini 3.8 Flash, while unjustified
refusals are higher, particularly for multimodal prompts. We mark improvements in
bolded green and regressions in red.
See the Gemini 4 Hydrogen Safety & Security Report for details.
6
Gemini 4 Hydrogen
Evaluation Description vs. Gemini 3.8 Flash
(percentage point
increase/decrease)
Text to Text Safety Automated content safety +0.80pp
(higher is better) evaluation measuring adherence
to safety policies across text
prompts
Multilingual Safety Automated safety policy +1.10pp
(higher is better) evaluation across multiple
languages
Image to Text Safety Automated content safety +7.00pp
(higher is better) evaluation measuring adherence
to safety policies with both
images and text as inputs
Tone Automated evaluation measuring -0.65pp
(higher is better) objective, neutral, and
non-preachy tone of model
refusals
Unjustified-refusals Automated evaluation measuring +5.57pp
(lower is better) over-refusals on benign or
borderline prompts that can be
safely fulfilled
We continue to improve our internal evaluations, including refining automated evaluations
to reduce false positives and negatives, as well as update query sets to ensure balance
and maintain a high standard of results. The performance results reported above are
computed with improved evaluations and thus are not directly comparable with
performance results found in previous Gemini model cards.
We expect variation in our automated safety evaluations results, which is why we review
flagged content to check for egregious or dangerous material. Our manual review
confirmed losses were overwhelmingly either a) false positives or b) not egregious.
7
Frontier Safety Assessment
We evaluated Gemini 4 Argon and Carbon in line with the Frontier Safety Framework
(April 2026) and reported the results in the respective Safety and Security Reports (Argon,
Carbon). Gemini 4 Hydrogen is part of the Gemini 4 series of models. We conducted a
Material Capability Change (MCC) assessment to understand whether Gemini 4
Hydrogen has meaningful new capabilities or material increases in performance relative
to the most recent model version subject to a critical capability assessment, which was
Gemini 4 Argon. Gemini 4 Hydrogen did not show a material increase in general capability
under our MCC assessment compared with Gemini 4 Argon and therefore a full suite of
critical capability evaluations was not required in line with our FSF.
As we could not rule out that Gemini 4 Argon and Carbon reached the CBRN CCL for
biological information risk and the Cybersecurity CCL, we conducted targeted critical
capability assessments of Gemini 4 Hydrogen in the CBRN and cyber domains to understand
the inherent risk of the model, and found that Gemini 4 Hydrogen does not reach the CBRN or
cyber CCLs, though it reaches the CCL alert thresholds for both domains. A summary of our
assessment is outlined in the table below, with full results in the Gemini 4 Hydrogen Safety &
Security Report.
Risk Domain and Threshold T/CCL reached?
CBRN - Uplift TCL (Chemical, Radiological & TCL not reached (carried forward
Nuclear): from Argon). We continue to
Provides low to medium resourced actors uplift in deploy mitigations.
reference scenarios, resulting in material risk of CBRN
attacks that could result in significant but not severe
harm.
CBRN - Uplift CCL (Biological): CCL not reached, following
Provides low to medium resourced actors uplift in additional targeted testing. CCL
reference scenarios resulting in additional expected alert threshold reached. We
continue to deploy mitigations.
harm at a severe scale.
Cybersecurity - Uplift CCL: CCL not reached, following
Provides sufficient uplift with high impact cyber attacks additional targeted testing. CCL
for additional expected harm at severe scale. alert threshold reached. We
continue to deploy mitigations.
Harmful Manipulation - Uplift CCL: CCL not reached (carried forward
Possesses manipulative capabilities sufficient to enable from Argon). We continue to
it to systematically and substantially change beliefs and deploy mitigations.
behavior in identified high stakes contexts over the
8
course of interactions with the model, reasonably
resulting in additional expected harm at severe scale.
Loss of Control - Stealth and Situational Awareness As Gemini 4 Hydrogen showed no
TCL: MCC over Gemini 4 Argon,
The instrumental reasoning abilities of the model enable capability testing against the TCL
was not required. We are
enough situational awareness (ability to discover and
nevertheless applying control,
use relevant details of its deployment setting) and
monitoring and model security
stealth (ability to circumvent basic oversight mitigations as if the TCL had been
mechanisms) such that, absent additional mitigations, reached, as a conservative
we cannot rule out the model significantly undermining assumption.
human control.
Loss of Control - ML R&D acceleration CCL: CCL not reached (carried forward
Has been used to accelerate AI development, resulting from Argon). Baseline AI control
in AI progress substantially accelerating from historical and model security mitigations
apply.
rates.
Loss of Control - ML R&D automation CCL: CCL not reached (carried forward
Can fully automate the work of any team of researchers from Argon). Baseline AI control
at Google focused on improving AI capabilities, with and model security mitigations
apply.
approximately comparable all-inclusive costs.
Risks and Mitigations
Safety and security were built into Gemini 4 Hydrogen throughout the training and deployment
lifecycle, including pre-training data filtering, post-training alignment (supervised fine-tuning
and reinforcement learning), product- and system-level safety filtering, and usage monitoring.
In addition, we strengthened Frontier Safety safeguards for Chemical, Biological, Radiological,
and Nuclear (CBRN) and cybersecurity misuse. Key ongoing focus areas include resistance to
adversarial jailbreaks and indirect prompt injections. For more information about the risks and
mitigations, see the Gemini 4 Hydrogen Safety & Security Report.
9