Model comparison
Gemini 3.8 Flash vs Gemini 4 Argon (restricted preview)
Two Google tiers compared on the figures that decide which one a workload actually needs.
Recorded profile verification October 1, 2026Individual provider fields may changeEvidence confidence: Insufficient data — see receipts below
Gemini 3.8 Flash
Balanced
Gemini 4 Argon (restricted preview)
Frontier
| Specification | Gemini 3.8 Flash | Gemini 4 Argon (restricted preview) |
|---|---|---|
| Provider | ||
| Provider | ||
| Tier | ||
| Tier | Balanced | Frontier |
| Context window | ||
| Context window | 1.05M | Not verifiedUnverified |
| Max output | ||
| Max output | 66K | Winner: 1M |
| Input / 1M tokens | ||
| Input / 1M tokens | $0.75 | Not verifiedUnverified |
| Output / 1M tokens | ||
| Output / 1M tokens | $3.75 | Not verifiedUnverified |
| Weights | ||
| Weights | Closed | Closed |
| Parameters | ||
| Parameters | Not verifiedUnverified | Not verifiedUnverified |
| Reasoning levels | ||
| Reasoning levels | low, medium, high | Not verifiedUnverified |
| Modalities | ||
| Modalities | text, image, video, audio, pdf | text, image, video |
| API model id | ||
| API model id | gemini-3.8-flash | Not verifiedUnverified |
| Released | ||
| Released | September 2, 2026 | September 30, 2026 |
| Artificial Analysis Intelligence Index [high] · source date Gemini 3.8 Flash: 2026-10-01 | ||
| Artificial Analysis Intelligence Index [high] · source date Gemini 3.8 Flash: 2026-10-01 | 41 | Not verifiedUnverified |
| Artificial Analysis Intelligence Index [high] · source date Gemini 4 Argon (restricted preview): 2026-10-01 | ||
| Artificial Analysis Intelligence Index [high] · source date Gemini 4 Argon (restricted preview): 2026-10-01 | Not verifiedUnverified | 53 |
| Terminal-Bench 2.1 · source date Gemini 3.8 Flash: 2026-09-02 | ||
| Terminal-Bench 2.1 · source date Gemini 3.8 Flash: 2026-09-02 | 89.4 | Not verifiedUnverified |
| HLE-Verified · source date Gemini 3.8 Flash: 2026-09-02 | ||
| HLE-Verified · source date Gemini 3.8 Flash: 2026-09-02 | 54.9 | Not verifiedUnverified |
| DeepSWE 1.1 · source date Gemini 4 Argon (restricted preview): 2026-10-01 | ||
| DeepSWE 1.1 · source date Gemini 4 Argon (restricted preview): 2026-10-01 | Not verifiedUnverified | 77.9 |
| AutomationBench · source date Gemini 4 Argon (restricted preview): 2026-10-01 | ||
| AutomationBench · source date Gemini 4 Argon (restricted preview): 2026-10-01 | Not verifiedUnverified | 51.3 |
| LVBench · source date Gemini 4 Argon (restricted preview): 2026-10-01 | ||
| LVBench · source date Gemini 4 Argon (restricted preview): 2026-10-01 | Not verifiedUnverified | 91.7 |
Dates beside benchmark scores refer to the linked source’s publication or capture date unless that source explicitly states when it measured the result. A shared URL alone does not establish comparable runs; unverified setups are marked. Prices are USD per million tokens at standard rates, excluding batch and caching discounts. Values we could not confirm from the provider are shown as “Not verified” rather than estimated.
Benchmark receipts
Where each score comes from, and how far it can be compared across models.
- Source date2026-10-01
- Source date2026-10-01
- Source date2026-10-01Google (vendor-reported; not in Datacurve’s September 22 board)
Agentic software engineering: pass@1 on real repo tasks with reported cost per completed task.
Directly comparable
- Source date2026-10-01Artificial Analysis (Gemini 4 Argon (high))
Composite index blending reasoning, knowledge, and coding evals into one 0–100 score.
Comparable with caveatOnly same-version scores are comparable. v4.3.2 was rebaselined (the top score fell from 66 on v4.1.1 to ~58), so v4.1.1 figures must not be compared with v4.3.2 figures — check the cited source and its date.
- Source date2026-10-01Artificial Analysis (Gemini 3.8 Flash (high))
Composite index blending reasoning, knowledge, and coding evals into one 0–100 score.
Comparable with caveatOnly same-version scores are comparable. v4.3.2 was rebaselined (the top score fell from 66 on v4.1.1 to ~58), so v4.1.1 figures must not be compared with v4.3.2 figures — check the cited source and its date.
- Source date2026-09-02Google official blog and methodology (vendor; full 1,811-item HLE-Verified set) (Full 1,811-item verified/revised set; differs from original HLE and AA text-only HLE)
Expert reasoning on the verified and revised Humanity's Last Exam question set.
UnknownA distinct question set from original HLE and Artificial Analysis's text-only HLE. Shared dataset and publisher alone do not establish matching model effort, tools, filtering and run configuration; no quality winner is inferred without that evidence.
- Source date2026-09-02Google primary evaluation table and methodology (vendor-reported) (Vendor evaluation; Terminus-2 harness; exact run date not published)
Agentic terminal work: multi-step tasks executed in a sandboxed shell environment.
Comparable with caveatNot comparable with Terminal-Bench 3.0 or 4.0 (different task sets) or v1; harness configuration (container, time limits) also shifts results.
Pricing tiers
Gemini 3.8 Flash: Intro Standard $0.75/$3.75 per MTok through December 31, 2026; $1.50/$7.50 from January 1, 2027. Output includes thinking tokens. Cached input $0.075, rising to $0.15; storage $0.50/MTok-hour, rising to $1. Batch/Flex input/output are $0.375/$1.875, rising to $0.75/$3.75. Priority $1.35/$6.75, rising to $2.70/$13.50. Grounding tools have separate charges and quotas; free-tier limits differ from paid.
Gemini 4 Argon (restricted preview): Google announced a future introductory API price of $2 input / $10 output per million tokens, with 95% off cached input. After the introductory period: $4/$20. These are announced launch rates; general API availability and the introductory end date are not confirmed. Current purchasable token rates are therefore left unverified.
Gemini 3.8 Flash
Gemini 3.8 Flash is Google's multimodal Flash model for long-horizon software engineering and agent workflows, with current introductory API pricing through December 31, 2026.
Best for
- Agentic coding at scale
- Mid-difficulty engineering
- Multimodal pipelines
Watch out
Intro rates and cache-storage prices double January 1, 2027. Google says 3.8 can use more tokens at higher effort; 3.7 remains supported for efficiency-first workloads. Minimal thinking errors. Computer use is Preview; native audio/image generation and Live API are unsupported. Restricted Flash Cyber is a separate endpoint with no public rate confirmed.
Gemini 4 Argon (restricted preview)
Google’s announced frontier model, initially restricted to trusted cyber defenders in the Fairwind program while broader safeguards and access are prepared.
Best for
- Long-horizon coding research
- Enterprise knowledge-work evaluation
Watch out
Restricted preview, not generally available. Google’s API model list does not yet publish a stable model ID or verified context limit. The 1M figure here is the announced maximum output, not a confirmed input context window.
Benchmark
DeepSWE 1.1 in context
Only Gemini 3.8 Flash has a published DeepSWE 1.1 result. It is shown against the wider field, with cost per completed task alongside the score.
Local leader
GPT-6 Astra [xhigh]
74.1%
Rows shown
28
Best source-matched Pass@1 per model
Snapshot date
2026-10-01
Mirrored from deepswe.datacurve.ai
Better is toward the top-right (higher pass rate, lower cost). X-axis is reversed to match DeepSWE’s public chart. v1.1 uses average cost, output tokens and steps. Legacy v1 transcriptions are excluded here until an exact primary receipt is retained.
| # | Model | Pass@1 | Cost / task | Tokens / task | Steps / task |
|---|---|---|---|---|---|
| 1 | GPT-6 Astra [xhigh] | 74.1% | $4.43 | 30k | 29 |
| 2 | Gemini 3.8 Flash [high] | 73.8% | $2.36 | 143k | 166 |
| 3 | Claude Opus 5 [max] | 73.6% | $11.84 | 118k | 99 |
| 4 | GPT-5.6 Sol [max] | 72.7% | $8.39 | 60k | 61 |
| 5 | Claude Fable 5 [xhigh] | 69.9% | $13.41 | 80k | 68 |
| 6 | GPT-5.6 Terra [max] | 69.6% | $4.95 | 72k | 76 |
| 7 | GLM 5.3 [max] | 69.0% | $3.99 | 80k | 124 |
| 8 | Kimi K3 [max] | 68.5% | $4.65 | 82k | 98 |
| 9 | Grok 4.6 [medium] | 67.5% | $3.45 | 50k | 70 |
| 10 | GPT-5.6 Luna [max] | 67.2% | $3.03 | 73k | 102 |
| 11 | GPT-5.5 [xhigh] | 67.0% | $7.23 | 46k | 82 |
| 12 | Gemini 3.7 Flash [medium] | 65.5% | $2.03 | 94k | 117 |
| 13 | GLM 5.3 Flash [max] | 63.4% | $0.48 | 73k | 123 |
| 14 | DeepSeek V4-Pro [max] | 62.8% | $0.24 | 106k | 155 |
| 15 | Claude Opus 4.8 [max] | 59.0% | $13.22 | 135k | 120 |
| 16 | Qwen3.8-Max [xhigh] | 57.5% | $3.73 | 95k | 111 |
| 17 | Muse Spark 1.2 [xhigh] | 54.9% | $3.70 | 99k | 101 |
| 18 | Claude Sonnet 5 [max] | 53.8% | $26.40 | 214k | 268 |
| 19 | Grok 4.5 [high] | 53.8% | $2.42 | 36k | 61 |
| 20 | DeepSeek V4-Flash [max] | 53.3% | $0.10 | 108k | 153 |
| 21 | Muse Spark 1.1 [xhigh] | 53.3% | $2.36 | 74k | 96 |
| 22 | GPT-5.4 [xhigh] | 51.8% | $5.65 | 71k | 70 |
| 23 | Gemini 3.6 Flash [high] | 46.7% | $4.42 | 96k | 117 |
| 24 | GLM 5.2 [max] | 43.8% | $3.92 | 78k | 129 |
| 25 | Gemini 3.5 Flash [high] | 36.1% | $3.45 | 76k | 105 |
| 26 | Kimi K2.7 Code | 30.5% | $2.82 | 59k | 149 |
| 27 | Claude Sonnet 4.6 [high] | 29.9% | $5.52 | 76k | 134 |
| 28 | Gemini 3.1 Pro [high] | 11.7% | $2.14 | 28k | 76 |
“Best per model” shows each model’s highest source-matched Pass@1 in this capture; switch to all effort levels to see every retained configuration. Efforts differ across models, so this view is not a comparison at matched effort. The source was generated September 22 and captured October 1. Small gaps may not be statistically meaningful — confirm on deepswe.datacurve.ai.
When the cheaper one wins
One of these rows is missing a verified output price, so this page will not name a cheaper winner. Run the model picker against the actual job instead of guessing.
Run the model pickerEvidence confidence: Insufficient data
How strong and complete the evidence behind this comparison is — not a prediction of which model is better. Hard limitations apply: missing pricing.
- Pricing incomplete — No verified token pricing for Gemini 4 Argon (restricted preview).
- 2/5 core specs verified on both sides — Not published for at least one side: context window, parameter count, reasoning levels.
- 1 benchmark setup aligned with differing scores — Matched source, source date, units, effort and recorded run details: Artificial Analysis Intelligence Index.
- Verified within the last 90 days — Older of the two catalog checks was 25 days ago.
- Both models carry source citations — Each side has at least two catalog sources on record.
Source receipts
Each catalog figure was checked against the provider or an independent second source on the date shown.
- Gemini 3.8 Flash: Google primary profile — whole-record source check October 1 (accessed 2026-10-01)
- Gemini 3.8 Flash: Gemini current API pricing and cache-storage tiers (accessed 2026-10-01)
- Gemini 3.8 Flash: Google — Gemini 3.8 Flash and 3.8 Flash Cyber (accessed 2026-09-03)
- Gemini 3.8 Flash: Gemini API pricing (gemini-3.8-flash $0.75/$3.75 intro) (accessed 2026-09-03)
- Gemini 3.8 Flash: Datacurve DeepSWE 1.1 — 3.8 Flash high 73.8%, $2.36/task; Google separately self-reports 73.7% (accessed 2026-09-03)
- Gemini 4 Argon (restricted preview): Google primary profile — whole-record source check October 1 (accessed 2026-10-01)
- Gemini 4 Argon (restricted preview): Gemini current API pricing and cache-storage tiers (accessed 2026-10-01)
- Gemini 4 Argon (restricted preview): Google announcement and availability (accessed 2026-10-01)
- Gemini 4 Argon (restricted preview): Gemini API model catalog (no general Argon endpoint) (accessed 2026-10-01)
- Gemini 4 Argon (restricted preview): Text Arena preview evaluation (accessed 2026-10-01)
Related comparisons
- Amazon Nova 2 Lite vs Gemini 3.8 Flash
- Amazon Nova 2 Pro vs Gemini 3.8 Flash
- Base 1 vs Gemini 3.8 Flash
- Claude Fable 5.1 vs Gemini 4 Argon (restricted preview)
- Claude Fable 5 vs Gemini 4 Argon (restricted preview)
- Claude Mythos 5.1 vs Gemini 4 Argon (restricted preview)
Diving deeper on one model? Gemini 3.8 Flash · Gemini 4 Argon (restricted preview)
Common questions
Gemini 3.8 Flash vs Gemini 4 Argon (restricted preview)
Answered from the verified figures on this page rather than general guidance.
Should I use Gemini 3.8 Flash or Gemini 4 Argon (restricted preview)?
Next step
Choosing between them
The verified figures that separate this pair, computed from the catalog rather than restated boilerplate.
- Published Artificial Analysis Intelligence Index results differ: Gemini 4 Argon (restricted preview) scores 53 to 41 on the same listed benchmark, source attribution, source date, effort, and run variant (2026-10-01).
- Positioning: Gemini 3.8 Flash sits in the balanced tier, Gemini 4 Argon (restricted preview) in the frontier tier. Compare accuracy, reliability, latency, and cost on representative tasks before choosing.
Choosing a harness rather than a model — Cursor, Copilot, Claude Code, Windsurf, or Lovable? Compare agentic harnesses · Latest releases.