Gemini 3.8 Flash pricing: real cost per million tokens
What Google's Gemini 3.8 Flash really costs across 5 plans from 3 providers: monthly fees, usable allowances and real per-token prices, against the official API list price and independent leaderboards.
Data updated · How real prices are computed
As of 2026-10-08, the cheapest way to use Gemini 3.8 Flash is Google AI Ultra 20x at a real cost of about $0.00387 per million tokens ($200/month for roughly 51.62B tokens a month). We track 5 plans and API channels that serve Gemini 3.8 Flash.
- Lowest real price
- $0.00387/MTokGoogle AI Ultra 20x
- API list price
- —
- Most monthly tokens
- 51.62BGoogle AI Ultra 20x
- Plans & APIs
- 55 subscriptions · 0 metered APIs
Plot Gemini 3.8 Flash on the price–capability chart →
Gemini 3.8 Flash plans and real prices
Real price = monthly fee ÷ tokens you can actually use in a month. “vs API” shows how many times cheaper than the official API list price (standard workload mix) the plan is. “Evidence” opens the full calculation and sources for that row.
| Plan | Provider | Billing | Monthly fee | Monthly tokens | Real $/MTok | vs API | Confidence | Data date | |
|---|---|---|---|---|---|---|---|---|---|
| Google AI Ultra 20x | Subscription | $200/mo | 51.62B | $0.00387 | — | Low | Evidence ↓ | ||
| Google AI Pro | Subscription | $19.99/mo | 2.581B | $0.00775 | — | High | Evidence ↓ | ||
| Google AI Ultra 5x | Subscription | $99.99/mo | 12.905B | $0.00775 | — | Low | Evidence ↓ | ||
| Droid Max | Factory | Subscription | $200/mo | 13.472B | $0.01485 | — | Medium | Evidence ↓ | |
| Command Code GOAT | Command Code | Subscription | $10/mo | 0.181B | $0.05513 | — | Medium | Evidence ↓ |
Behind each number
For every plan: the data date, the exact calculation, the quota basis, how leaderboard scores were mapped, and the original adoption evidence with its sources.
Google AI Ultra 20xGoogle$0.00387427/MTok
- Plan
- Google AI Ultra 20x · Google
- Data date
- 2026-09-12 – 2026-09-27 · derived; uses the data date of Google AI Pro · gemini-3.8-flash
- Real price / MTok
- $0.00387427
- Monthly tokens
- 51,620,000,000 (51.62B)
- Quota confidence
- Low
- Calculation
- $200 / month ÷ 51,620,000,000 tokens × 1,000,000 ≈ $0.00387427 / MTok
- Quota basis
Low-cache workload: 85% cache reads / 14.5% input / 0.5% output
- Benchmark mapping
- Artificial Analysis Intelligence Index v4.3: Medium (3 configurations)
Terminal-Bench 4.0: Low (1 configuration)
Terminal-Bench 4.0 (AA): Low (3 configurations)
Code Arena | WebDev Overall: Low (1 configuration)
Agent Arena · Overall (Ranking / Pareto): Low (1 configuration)
Artificial Analysis Coding Agent Index v1.5: Low (1 configuration)
OpenDesign Arena · Quality ranking: Low (1 configuration)
DeepSWE v1.1: Low (2 configurations)Mapping confidence: how directly each score refers to the model served here.
Adoption evidence · source and rationale (translated from the Chinese record)
Official: Ultra $200 = 20× Pro token worth (antigravity.google/blog 2026-05-19)
Added 51.62B: Pro adopted value × official worth multiplier 20; not an independent measurement
Google AI ProGoogle$0.00774506/MTok
- Plan
- Google AI Pro · Google
- Data date
- 2026-09-12 – 2026-09-27 · sampled
- Real price / MTok
- $0.00774506
- Monthly tokens
- 2,581,000,000 (2.581B)
- Quota confidence
- High
- Calculation
- $19.99 / month ÷ 2,581,000,000 tokens × 1,000,000 ≈ $0.00774506 / MTok
- Quota basis
Low-cache workload: 85% cache reads / 14.5% input / 0.5% output
- Benchmark mapping
- Artificial Analysis Intelligence Index v4.3: Medium (3 configurations)
Terminal-Bench 4.0: Low (1 configuration)
Terminal-Bench 4.0 (AA): Low (3 configurations)
Code Arena | WebDev Overall: Low (1 configuration)
Agent Arena · Overall (Ranking / Pareto): Low (1 configuration)
Artificial Analysis Coding Agent Index v1.5: Low (1 configuration)
OpenDesign Arena · Quality ranking: Low (1 configuration)
DeepSWE v1.1: Low (2 configurations)Mapping confidence: how directly each score refers to the model served here.
Adoption evidence · source and rationale (translated from the Chinese record)
issue #54 (NTRYourWaifu): agy /quota polling the weekly-quota % every 5 minutes across three weekly windows, counting only segments where 3.8 Flash is ≥95%: 853.43M raw = +165.77pp; community-issues-round1-2026-09-29.json; maintainer's local measurement: the whole B segment 55.343M raw (cache 45.69M / in 9.26M / out 0.40M) = weekly bar +9.88%; gemini-weekly-round7-2026-09-21.json; workload-conversion-round1-2026-09-30.json
2.070B→2.581B (2026-09-30 decision: uniform workload-normalization for measured samples with token breakdown, low-cache tier): the two samples' worth merged per percentage point (local B segment cache 45.69M / in 9.26M / out 0.40M = $11.8626 + #54 cache 695.70M / in 146.25M / out 11.48M = $204.915) ÷175.65pp = $1.23415/pp → $123.42/week ×4 weeks ÷ low-cache blended price $0.19125/MTok; the measured 82~84% cache belongs to the low-cache family, so the standard-tier basis 4.478B was not adopted; the raw-basis 20.7 is kept for comparison. Original 2026-09-29 adoption decision (raw basis) — 2.241B→2.070B, 2026-09-29 decision: issue #54 (agy /quota weekly-quota % every 5 minutes, three weekly windows, only segments where 3.8 Flash exceeds 95%) 853.43M raw = +165.77 weekly percentage points (20.59 alone), merged per percentage point with the maintainer's local sample 55.343M = +9.88%: (55.343+853.43)÷(9.88+165.77) = 5.1738M/pp ×100×4 weeks = 2.070B raw. Both samples have ~82.6% cache hits. Worth per 1% is about $1.19~1.25, independent of effort. Evidence: community-issues-round1-2026-09-29.json; the original sample is in the gemini-weekly-round evidence.
community-issues-round1-2026-09-29.jsongemini-weekly-round7-2026-09-21.jsonworkload-conversion-round1-2026-09-30.jsonBrowse the public evidence archive
Google AI Ultra 5xGoogle$0.00774816/MTok
- Plan
- Google AI Ultra 5x · Google
- Data date
- 2026-09-12 – 2026-09-27 · derived; uses the data date of Google AI Pro · gemini-3.8-flash
- Real price / MTok
- $0.00774816
- Monthly tokens
- 12,905,000,000 (12.905B)
- Quota confidence
- Low
- Calculation
- $99.99 / month ÷ 12,905,000,000 tokens × 1,000,000 ≈ $0.00774816 / MTok
- Quota basis
Low-cache workload: 85% cache reads / 14.5% input / 0.5% output
- Benchmark mapping
- Artificial Analysis Intelligence Index v4.3: Medium (3 configurations)
Terminal-Bench 4.0: Low (1 configuration)
Terminal-Bench 4.0 (AA): Low (3 configurations)
Code Arena | WebDev Overall: Low (1 configuration)
Agent Arena · Overall (Ranking / Pareto): Low (1 configuration)
Artificial Analysis Coding Agent Index v1.5: Low (1 configuration)
OpenDesign Arena · Quality ranking: Low (1 configuration)
DeepSWE v1.1: Low (2 configurations)Mapping confidence: how directly each score refers to the model served here.
Adoption evidence · source and rationale (translated from the Chinese record)
Official: Ultra $100 = 5× Pro token worth (antigravity.google/blog 2026-05-19)
Added 12.905B: Pro adopted value × official worth multiplier 5; LLMDevs' Ultra ~5.0B/week is same-magnitude corroboration; not an independent measurement
Droid MaxFactory$0.0148456/MTok
- Data date
- 2026-09-29 · derived; uses the data date of Droid Max · claude-opus-5.5
- Real price / MTok
- $0.0148456
- Monthly tokens
- 13,472,000,000 (13.472B)
- Quota confidence
- Medium
- Calculation
- $200 / month ÷ 13,472,000,000 tokens × 1,000,000 ≈ $0.0148456 / MTok
- Quota basis
Anthropic workload: 97% cache reads / 2.5% cache writes / 0.5% output
- Benchmark mapping
- Artificial Analysis Intelligence Index v4.3: Medium (3 configurations)
Terminal-Bench 4.0: Low (1 configuration)
Terminal-Bench 4.0 (AA): Low (3 configurations)
Code Arena | WebDev Overall: Low (1 configuration)
Agent Arena · Overall (Ranking / Pareto): Low (1 configuration)
Artificial Analysis Coding Agent Index v1.5: Low (1 configuration)
OpenDesign Arena · Quality ranking: Low (1 configuration)
DeepSWE v1.1: Low (2 configurations)Mapping confidence: how directly each score refers to the model served here.
Adoption evidence · source and rationale (translated from the Chinese record)
Derived from same-plan claude-opus-5.5 2.526B × 5.333333333333334
Factory official multipliers Opus 5.5 1.6× / gemini-3.8-flash 0.3×: the same Standard Usage pool scaled by multiplier — not a measurement of this model; inherits the Opus 5.5 row's Anthropic-tier workload, not recomputed for each model's own price gap; docs.factory.ai/docs/models; droid-model-multipliers-2026-09-29.json; 0.3× is a promotional multiplier reverting to 0.6× after 2027-01-01 (then ≈2.66667× Opus); must re-verify at expiry
droid-model-multipliers-2026-09-29.jsonBrowse the public evidence archive
Command Code GOATCommand Code$0.0551268/MTok
- Data date
- 2026-09-30 · official source date
- Real price / MTok
- $0.0551268
- Monthly tokens
- 181,400,000 (0.181B)
- Quota confidence
- Medium
- Calculation
- $10 / month ÷ 181,400,000 tokens × 1,000,000 ≈ $0.0551268 / MTok
- Quota basis
Standard workload: 97% cache reads / 2.5% input / 0.5% output
- Benchmark mapping
- Artificial Analysis Intelligence Index v4.3: Medium (3 configurations)
Terminal-Bench 4.0: Low (1 configuration)
Terminal-Bench 4.0 (AA): Low (3 configurations)
Code Arena | WebDev Overall: Low (1 configuration)
Agent Arena · Overall (Ranking / Pareto): Low (1 configuration)
Artificial Analysis Coding Agent Index v1.5: Low (1 configuration)
OpenDesign Arena · Quality ranking: Low (1 configuration)
DeepSWE v1.1: Low (2 configurations)Mapping confidence: how directly each score refers to the model served here.
Adoption evidence · source and rationale (translated from the Chinese record)
https://commandcode.ai/docs/plans/goat official per-model allowance and three-part pricing; https://commandcode.ai/pricing; $10->$70 credits; code-subscriptions-round1-2026-09-06.json; goat-opencode-catalogs-round3-2026-09-30.json
Added 181.4M: min(shared monthly pool $70, model allowance $40) ÷ project standard-workload weighted price; official three-part pricing. Processing fee ignored; allowances of different models in the same plan are not additive; no usage-panel token+% screenshot — computed from official absolute credits and the price list
https://commandcode.ai/docs/plans/goathttps://commandcode.ai/pricing;code-subscriptions-round1-2026-09-06.jsongoat-opencode-catalogs-round3-2026-09-30.jsonBrowse the public evidence archive
Gemini 3.8 Flash benchmark scores
Highest archived configuration per leaderboard; ranks are among all models each board archives in this dataset. Scores are never mixed across boards.
| Leaderboard | Metric | Score | Rank | Configuration | Effort | Harness | Configs | Snapshot |
|---|---|---|---|---|---|---|---|---|
| Artificial Analysis Intelligence Index v4.3 | Intelligence Index | 40.93 | #21 / 60 | Gemini 3.8 Flash (high) | High | — | 3 | 2026-10-01 |
| Terminal-Bench 4.0 | Resolution Rate % | 19.09 | #14 / 18 | mini-SWE-agent - Gemini 3.8 Flash (high) | High | mini-SWE-agent | 1 | 2026-09-21 |
| Terminal-Bench 4.0 (AA) | Resolution Rate % | 19.7 | #22 / 53 | Artificial Analysis - Gemini 3.8 Flash (medium) | Medium | Artificial Analysis | 3 | 2026-10-01 |
| Code Arena | WebDev Overall | Arena Score | 1,583 | #24 / 54 | gemini-3.8-flash-high | High | — | 1 | 2026-09-30 |
| Agent Arena · Overall (Ranking / Pareto) | Net Improvement % | 2.96 | #17 / 39 | Gemini 3.8 Flash (High) | High | — | 1 | 2026-09-30 |
| Artificial Analysis Coding Agent Index v1.5 | Coding Agent Index | 41.86 | #17 / 19 | Antigravity SDK - Gemini 3.8 Flash (high) | High | Antigravity SDK v0.1.12 | 1 | 2026-10-01 |
| OpenDesign Arena · Quality ranking | Average task score | 68.9 | #14 / 16 | Gemini 3.8 Flash | — | OpenDesign | 1 | 2026-10-01 |
| DeepSWE v1.1 | Pass@1 % | 73.83 | #3 / 33 | mini-swe-agent - gemini-3-8-flash (high) | High | mini-swe-agent | 2 | 2026-09-03 |
Every archived configuration
Artificial Analysis Intelligence Index v4.3 3 configurations
| Configuration | Score | Score interval | Harness / effort / mode | Source task cost (mean / median) | Checked | Links |
|---|---|---|---|---|---|---|
| Gemini 3.8 Flash (high) | 40.9262 | — | — / High / — | — / — | 2026-10-01 | Source · Archive |
| Gemini 3.8 Flash (medium) | 39.774 | — | — / Medium / — | — / — | 2026-10-01 | Source · Archive |
| Gemini 3.8 Flash (low) | 33.4548 | — | — / Low / — | — / — | 2026-10-01 | Source · Archive |
Terminal-Bench 4.0 1 configuration
| Configuration | Score | Score interval | Harness / effort / mode | Source task cost (mean / median) | Checked | Links |
|---|---|---|---|---|---|---|
| mini-SWE-agent - Gemini 3.8 Flash (high)Official Terminal-Bench 4.0 leaderboard row; score belongs to this harness × model × effort configuration. | 19.09 | 15.73 – 22.45 | mini-SWE-agent / High / — | — / — | 2026-10-01 | Source · Archive |
Terminal-Bench 4.0 (AA) 3 configurations
| Configuration | Score | Score interval | Harness / effort / mode | Source task cost (mean / median) | Checked | Links |
|---|---|---|---|---|---|---|
| Artificial Analysis - Gemini 3.8 Flash (medium)AA's own-harness independent measurement of Terminal-Bench 4.0 (the terminalBench40 field in the models-board payload, fraction×100); scored separately from the official tbench.ai board. | 19.697 | — | Artificial Analysis / Medium / — | — / — | 2026-10-01 | Source · Archive |
| Artificial Analysis - Gemini 3.8 Flash (high)AA's own-harness independent measurement of Terminal-Bench 4.0 (the terminalBench40 field in the models-board payload, fraction×100); scored separately from the official tbench.ai board. | 19.697 | — | Artificial Analysis / High / — | — / — | 2026-10-01 | Source · Archive |
| Artificial Analysis - Gemini 3.8 Flash (low)AA's own-harness independent measurement of Terminal-Bench 4.0 (the terminalBench40 field in the models-board payload, fraction×100); scored separately from the official tbench.ai board. | 10.101 | — | Artificial Analysis / Low / — | — / — | 2026-10-01 | Source · Archive |
Code Arena | WebDev Overall 1 configuration
Agent Arena · Overall (Ranking / Pareto) 1 configuration
Artificial Analysis Coding Agent Index v1.5 1 configuration
| Configuration | Score | Score interval | Harness / effort / mode | Source task cost (mean / median) | Checked | Links |
|---|---|---|---|---|---|---|
| Antigravity SDK - Gemini 3.8 Flash (high)Score belongs to this named agent/harness × model configuration; it is not a harness-independent model measurement. indexScore stored as percent (raw fraction × 100). | 41.8632 | — | Antigravity SDK v0.1.12 / High / — | $2.467 / — | 2026-10-01 | Source · Archive |
OpenDesign Arena · Quality ranking 1 configuration
DeepSWE v1.1 2 configurations
| Configuration | Score | Score interval | Harness / effort / mode | Source task cost (mean / median) | Checked | Links |
|---|---|---|---|---|---|---|
| mini-swe-agent - gemini-3-8-flash (high) | 73.8255 | 72.408 – 75.243 | mini-swe-agent / High / — | $2.362 / $2.11 | 2026-09-12 | Source · Archive |
| mini-swe-agent - gemini-3-8-flash (medium) | 71.0177 | 68.737 – 73.299 | mini-swe-agent / Medium / — | $1.967 / $1.734 | 2026-09-12 | Source · Archive |
These are reference mappings, not benchmarks of a subscription or API channel. Harness and quota-measurement effort alignment are unverified. Source task costs are not subscription task costs.
- Exact served-model reference; quota-measurement effort is unverified.
- Exact served-model reference only; product harness and quota-measurement effort are unverified. Not a benchmark measurement of this subscription or API channel.
Frequently asked questions
How much does Gemini 3.8 Flash really cost per million tokens?
Measured as monthly fee ÷ tokens you can actually use, Gemini 3.8 Flash ranges from $0.00387 (Google AI Ultra 20x) to $0.05513 (Command Code GOAT) per million tokens.
Which subscription gives the most Gemini 3.8 Flash usage?
Google AI Ultra 20x allows roughly 51.62B tokens of Gemini 3.8 Flash per month for $200.
How does Gemini 3.8 Flash score on independent leaderboards?
Gemini 3.8 Flash has scores on 8 leaderboards; its strongest placement is 73.83 (Pass@1 %) on DeepSWE v1.1, #3 of 33 models tracked.
Who makes Gemini 3.8 Flash, and where can I use it?
Gemini 3.8 Flash (model ID gemini-3.8-flash) is made by Google and is available through Command Code, Factory, Google.
More Google models
| Model | Lowest real $/MTok | Cheapest plan | Plans |
|---|---|---|---|
| Gemini 3.6 Flash | $0.00803 | Google AI Pro | 1 |
| Gemini 3.7 Flash | $0.01485 | Droid Max | 2 |