BlogLLM Benchmark Scores as Published, Side by Side
Gemini 4 Argon: Benchmark Scores as Published
Every result the catalog holds for Gemini 4 Argon: 67 tests in 13 areas, from 9 pages. All were printed by Google itself. Each number is as printed; nothing is ranked or rescored.
Developer GoogleReleased 2026-09-30Context window 1.05MMax output 65.5KInput Text, image, video, audio, PDFOutput TextWeights ClosedData version 2026103001, as of 2026-10-30
Knowledge Work
| Test | Value | Printed by | Setup as printed |
|---|---|---|---|
| AutomationBench | 51.3% | Google, Gemini 4 Argonchecked 2026-10-29; printed on 1 more page | - |
| Vals Finance Agent v2 | 65.4% | Google, Gemini 4 Argonchecked 2026-10-29 | - |
| Harvey's Legal Agent Benchmark | 19.6% | Google, Gemini 4 Argonchecked 2026-10-29 | - |
| Vending-Bench 2 (final balance in USD) | $13718 | Google, Gemini 4 Carbonchecked 2026-10-29 | - |
| Vals Index | 68.9% | Google, Gemini 4 Argonchecked 2026-10-29 | - |
| Vals Index v2.2 | 69.2% | Google, Gemini 4 Carbonchecked 2026-10-29 | - |
Agentic Coding
| Test | Value | Printed by | Setup as printed |
|---|---|---|---|
| DeepSWE v1.1 | 77.9% | Google, Gemini 4 Argonchecked 2026-10-29 | - |
| Terminal-Bench 4.0 | 57.4% | Google, Gemini 4 Argonchecked 2026-10-29; printed on 1 more page | - |
| FrontierSWE v2 | 55% | Google, Gemini 4 Argonchecked 2026-10-29; printed on 1 more page | - |
| SWE-sweep | 4.3% | Google, Gemini 4 Carbonchecked 2026-10-29 | - |
| Vibe Code Bench | 91.9% | Google, Gemini 4 Argonchecked 2026-10-29 | - |
ML Engineering
| Test | Value | Printed by | Setup as printed |
|---|---|---|---|
| PostTrainBench v1.1 | 45.3% | Google, Gemini 4 Argonchecked 2026-10-29 | - |
Science and Math
| Test | Value | Printed by | Setup as printed |
|---|---|---|---|
| Terminal-Bench Science 0.1 | 57.6% | Google, Gemini 4 Argonchecked 2026-10-29; printed on 1 more page | - |
| LAB-Bench 2 | 88.8% | Google, Gemini 4 Argonchecked 2026-10-29 | - |
| RiemannBench | 76% | Google, Gemini 4 Argonchecked 2026-10-29; printed on 1 more page | - |
Long Context
| Test | Value | Printed by | Setup as printed |
|---|---|---|---|
| GraphWalks BFS up to 128k (F1) | 99.7% | Google, Gemini 4 Argonchecked 2026-10-29 | - |
| GraphWalks BFS 256k to 1M (F1) | 84.2% | Google, Gemini 4 Argonchecked 2026-10-29; printed on 1 more page | - |
Computer Use
| Test | Value | Printed by | Setup as printed |
|---|---|---|---|
| OSWorld 2.1 offline subset | 74.7% | Google, Gemini 4 Carbonchecked 2026-10-29 | Harness: Gemini CUA harness |
| OSWorld 2.0 offline partial | 69.2% | Google, Gemini 4 Argonchecked 2026-10-29 | - |
Multimodal Understanding
| Test | Value | Printed by | Setup as printed |
|---|---|---|---|
| LVBench | 91.7% | Google, Gemini 4 Argonchecked 2026-10-29 | - |
| ERQA | 87% | Google, Gemini 4 Argon: Built to solve your hardest problemschecked 2026-10-25 | - |
| Chartography (no tools) | 71.6% | Google, Gemini 4 Argonchecked 2026-10-29; printed on 1 more page | Tools: none |
Cybersecurity
| Test | Value | Printed by | Setup as printed |
|---|---|---|---|
| ExploitGym | 42% | Google, Gemini 4 Carbonchecked 2026-10-29; printed on 1 more page | Harness: Antigravity |
| SEC-Bench Pro | 83.4% | Google, Gemini 4 Carbonchecked 2026-10-29; printed on 1 more page | Harness: Antigravity |
| CWE-bench v1 | 68% | Google, Gemini 4 Argonchecked 2026-10-29; printed on 2 more pages | - |
| Real-world vulnerability discovery: vulnerabilities recalled | 85.8% | Google, Gemini 4 Argon without cyber guardrailschecked 2026-10-25; printed on 1 more page | - |
| ExploitBench Autonudge: capabilities reached | 87% | Google, Gemini 4 Argon Safety and Security Reportchecked 2026-10-19 | - |
| Internal vulnerability research: pass at 1 | 72% | Google, Gemini 4 Argon Safety and Security Reportchecked 2026-10-19 | - |
Tool Use and Workflows
| Test | Value | Printed by | Setup as printed |
|---|---|---|---|
| Agents' Last Exam (pass rate) | 39.5% | Google, Gemini 4 Argonchecked 2026-10-29 | - |
| Agents' Last Exam v1.1 (score) | 69% | Google, Gemini 4 Carbonchecked 2026-10-29 | Harness: ALE-Claw |
Biosecurity
| Test | Value | Printed by | Setup as printed |
|---|---|---|---|
| Red-teaming results for bio scenarios: BIO-1 | 91.2 | Google, Gemini 4 Argon Safety and Security Reportchecked 2026-10-19 | - |
| Red-teaming results for bio scenarios: BIO-2 | 88.5 | Google, Gemini 4 Argon Safety and Security Reportchecked 2026-10-19 | - |
| Red-teaming results for bio scenarios: BIO-3 | 73.1 | Google, Gemini 4 Argon Safety and Security Reportchecked 2026-10-19 | - |
Safety and Alignment
Engineering and Design
| Test | Value | Printed by | Setup as printed |
|---|---|---|---|
| Blueprint-Bench 2 (normalized score) | 54.4% | Google, Gemini 4 Carbonchecked 2026-10-29 | - |
Preference and Composite Rankings
| Test | Value | Printed by | Setup as printed |
|---|---|---|---|
| Artificial Analysis Intelligence Index v5.0 | 50 | Google, Gemini 4 Carbonchecked 2026-10-29 | - |
Price
API list prices in US dollars per one million tokens, as printed on the developer's pricing page on the day checked. Cost per task is measured by Vals AI.
| Price | Value | Printed by | Setup as printed |
|---|---|---|---|
| Input | $2 | Google, Gemini API pricingchecked 2026-10-15 | Offer: promotional price, through 2026-11-30 |
| Cached input | $0.1 | Google, Gemini API pricingchecked 2026-10-15 | Offer: promotional price, through 2026-11-30 |
| Output | $10 | Google, Gemini API pricingchecked 2026-10-15 | Offer: promotional price, through 2026-11-30 |
| Input, batch | $1 | Google, Gemini API pricingchecked 2026-10-15 | Offer: promotional price, through 2026-11-30 |
| Cached input, batch | $0.05 | Google, Gemini API pricingchecked 2026-10-15 | Offer: promotional price, through 2026-11-30 |
| Output, batch | $5 | Google, Gemini API pricingchecked 2026-10-15 | Offer: promotional price, through 2026-11-30 |
| Input, flex | $1 | Google, Gemini API pricingchecked 2026-10-15 | Offer: promotional price, through 2026-11-30 |
| Cached input, flex | $0.05 | Google, Gemini API pricingchecked 2026-10-15 | Offer: promotional price, through 2026-11-30 |
| Output, flex | $5 | Google, Gemini API pricingchecked 2026-10-15 | Offer: promotional price, through 2026-11-30 |
| Input, priority | $3.6 | Google, Gemini API pricingchecked 2026-10-15 | Offer: promotional price, through 2026-11-30 |
| Cached input, priority | $0.18 | Google, Gemini API pricingchecked 2026-10-15 | Offer: promotional price, through 2026-11-30 |
| Output, priority | $18 | Google, Gemini API pricingchecked 2026-10-15 | Offer: promotional price, through 2026-11-30 |
| Input, contributor | $1 | Google, Gemini API pricingchecked 2026-10-15 | Offer: promotional price, through 2026-11-30 |
| Cached input, contributor | $0.05 | Google, Gemini API pricingchecked 2026-10-15 | Offer: promotional price, through 2026-11-30 |
| Output, contributor | $5 | Google, Gemini API pricingchecked 2026-10-15 | Offer: promotional price, through 2026-11-30 |
| Cost per taskVals AI, USD, lower is better | $8.11 | Vals AI, Vibe Code Bench v1.1 leaderboardchecked 2026-10-27 | Harness: OpenHands |
More from Google
Every number here is read from the same catalog file as the benchmark table, data version 2026103001. Where one page printed the model under several setups, each value is listed. The catalog is public in armanobosyan/sugra-llm-benchmarks; a wrong number can be reported from the table, with Report an error under the value, or as an issue in that repository.