GoogleProprietary model
Gemini 4 Argon benchmarks
- General rankToo few tasks3 non-coding tasks ranked5 needed for a rank
- Speed#7 / 76,714 msMedian latency, p50Hosted API only
- Cost#9 / 10$31Per 1,000 decisionsHosted price
- Scenario judge#1 / 10 tied84.0%Best task · Accuracyvs frontier LLMs
Comparison summary
- Strongest
- Judging and evals
- Leads 1 of 3 (33%)
A language model (LLM): it reads a prompt and writes text back, so the prompt sets the task. Strongest at judging.
Specifications
- Lab
- Weights
- Proprietary
- Parameters
- Not disclosed
- Licence
- Proprietary
Judging and evals
Share of tasks led · #5 of 10
- Claude Opus 5.5Leads 3 of 3
- Gemini 4 HydrogenLeads 2 of 3
- GPT-6.1 SolLeads 2 of 3
- OpenAI Decisions APILeads 4 of 6
- Gemini 4 ArgonLeads 1 of 3
- Claude Sonnet 5.5Leads 1 of 3
- Gemini 3.8 FlashLeads 1 of 3
- JevLeads 1 of 6
See the 3 tasks
LLM-as-a-judge
Tier 2 of 5#3 / 10Accuracy, % · Higher is better
- GPT-6.1 Sol92.6, interval 89.1 to 95.8, tied for the lead
- Claude Opus 5.589.1, interval 83.9 to 93.9, tied for the lead
- Gemini 4 Hydrogen87.0, interval 83.2 to 90.6, tier 2
- Gemini 4 Argon84.4, interval 79.7 to 89.1, tier 2
- Claude Sonnet 5.583.5, interval 78.4 to 88.5, tier 2
- GPT-6 Luna82.6, interval 77.5 to 87.8, tier 2
- Gemini 3.8 Flash80.9, interval 76.2 to 85.5, tier 3
- Claude Haiku 5.573.9, interval 67.9 to 80.3, tier 3
Scenario judge
Tied for the lead#1 / 10Accuracy, % · Higher is better
- Claude Opus 5.586.8, interval 80.2 to 92.3, tied for the lead
- Gemini 4 Hydrogen86.1, interval 79.8 to 91.6, tied for the lead
- Gemini 4 Argon84.0, interval 77.7 to 89.6, tied for the lead
- Gemini 3.8 Flash83.3, interval 76.8 to 89.1, tied for the lead
- OpenAI Decisions API83.3, interval 76.1 to 89.9, tied for the lead
- Claude Sonnet 5.581.3, interval 74.1 to 87.4, tier 2
- GPT-6.1 Sol81.3, interval 74.1 to 87.3, tier 2
- GPT-6 Luna79.9, interval 72.8 to 86.2, tier 2
Search
Tier 2 of 4#5 / 10Accuracy, % · Higher is better
- Claude Opus 5.591.5, interval 87.6 to 94.8, tied for the lead
- Gemini 4 Hydrogen90.0, interval 86.9 to 92.8, tied for the lead
- GPT-6.1 Sol89.6, interval 85.6 to 93.1, tied for the lead
- Claude Sonnet 5.589.1, interval 85.8 to 92.1, tied for the lead
- Gemini 3.8 Flash85.0, interval 81.3 to 88.4, tier 2
- Gemini 4 Argon81.7, interval 77.9 to 85.4, tier 2
- Claude Haiku 5.581.3, interval 76.9 to 85.4, tier 2
- GPT-6 Luna75.2, interval 70.7 to 79.6, tier 3
Speed
Median latency, p50 · Hosted API only · lower is better
- OpenAI Decisions API226 ms
- Jev318 ms
- Claude Haiku 5.51,966 ms
- GPT-6 Luna3,085 ms
- Gemini 3.8 Flash5,493 ms
- Gemini 4 Hydrogen6,516 ms
- Gemini 4 Argon6,714 ms
Cost
Per 1,000 decisions, hosted price · lower is better
- Jev$0.43
- OpenAI Decisions API$1.23
- GPT-6 Luna$1.42
- Claude Haiku 5.5$2.17
- Gemini 3.8 Flash$15
- GPT-6.1 Sol~$28
- Claude Sonnet 5.5~$30
- Gemini 4 Argon$31
How we count
A model leads a task when it is in the task's leading tie tier. Each task counts inside its own benchmark, against that benchmark's models; no score is averaged. An area chart shows the models ranked on at least 2 of the area's tasks in a benchmark this model is in too. The overall standing counts every non-coding tasks and needs 5 for a rank. Latency is compared only on one hardware tier.
Gemini 4 ArgonOther modelsWhisker: 95% interval
