Google / Gemini
Gemini 3.8 Flash
Version gemini-3-8-flash · Released 2026-09-02 · Available
Compare this model ↗01 / EVIDENCE
Benchmark breakdown
| Benchmark | Version | Method | Configuration | Result | Test date | Verified | Source type | Source |
|---|---|---|---|---|---|---|---|---|
| ARC-AGI-3 Semi-Private Standard harness best effort | 3 Semi-Private | ARC Prize Standard harness; best published effort per model | — | 10.3663 % | 2026-09-29 | 2026-09-29 | benchmark maintainer | View source ↗ |
| Humanity's Last Exam AA Index v4.3.2 (Opus Max fallback, Astra Max, Gemini High) | AA Index 4.3.2 | Gemini High | — | 48 % | 2026-09-29 | 2026-09-29 | independent lab | View source ↗ |
| AutomationBench-AA Index v4.3.2 (Opus Max fallback, Astra Max, Gemini High) | AA Index 4.3.2 | Gemini High | — | 60 % | 2026-09-29 | 2026-09-29 | independent lab | View source ↗ |
| LiveBench code_completion | 2026-06-25 | Public table; published model effort variant | gemini-3.8-flash-high | 71.739 % | not published | 2026-09-30 | benchmark maintainer | View source ↗ |
| LiveBench code_generation | 2026-06-25 | Public table; published model effort variant | gemini-3.8-flash-high | 73.239 % | not published | 2026-09-30 | benchmark maintainer | View source ↗ |
| LiveBench javascript | 2026-06-25 | Public table; published model effort variant | gemini-3.8-flash-high | 72.727 % | not published | 2026-09-30 | benchmark maintainer | View source ↗ |
| LiveBench AMPS_Hard | 2026-06-25 | Public table; published model effort variant | gemini-3.8-flash-high | 99 % | not published | 2026-09-30 | benchmark maintainer | View source ↗ |
| LiveBench olympiad | 2026-06-25 | Public table; published model effort variant | gemini-3.8-flash-high | 92.159 % | not published | 2026-09-30 | benchmark maintainer | View source ↗ |
| LiveBench theory_of_mind | 2026-06-25 | Public table; published model effort variant | gemini-3.8-flash-high | 76.923 % | not published | 2026-09-30 | benchmark maintainer | View source ↗ |
| LiveBench zebra_puzzle | 2026-06-25 | Public table; published model effort variant | gemini-3.8-flash-high | 98.25 % | not published | 2026-09-30 | benchmark maintainer | View source ↗ |
02 / COST
Price & speed
Input / 1M tokens $0.75
Output / 1M tokens $3.75
Cached input / 1M —
Batch input / 1M —
Batch output / 1M —
Speed No comparable observation yet
Price verified 2026-09-29
Price scope Introductory rate through 2026-12-31
Price source Official pricing ↗
02A / PERFORMANCE
Measured speed
No sourced speed measurement at a stated effort level yet.
03 / PROFILE
Capabilities
04 / CONTEXT
Strengths & limitations
Coverage: 73% · Moderate confidence · 10 observations · 3 source domains (3 independent).
05 / HISTORY
Version & price history
Version gemini-3-8-flash · released 2026-09-02.
2026-09-29: $0.75 input / $3.75 output per 1M tokens · Price source ↗
06 / SOURCES
Source records
- Gemini 3.8 Flash official documentation ↗ · verified 2026-09-29
- Google official pricing ↗ · verified 2026-09-29
DATED EVIDENCE
History
Standard API price history
- 2026-09-29: 1.5 USD / 1M · 75% input / 25% output USD per 1M · source ↗
One observation; no trend can be inferred.
ARC-AGI-3 Semi-Private Standard harness best effort / 3 Semi-Private / %
- 2026-09-29: 10.366 result · ARC Prize Standard harness; best published effort per model · source ↗
One observation; no trend can be inferred.
Humanity's Last Exam AA Index v4.3.2 (Opus Max fallback, Astra Max, Gemini High) / AA Index 4.3.2 / %
- 2026-09-29: 48 result · Gemini High · source ↗
One observation; no trend can be inferred.
AutomationBench-AA Index v4.3.2 (Opus Max fallback, Astra Max, Gemini High) / AA Index 4.3.2 / %
- 2026-09-29: 60 result · Gemini High · source ↗
One observation; no trend can be inferred.
LiveBench code_completion / 2026-06-25 / %
- 2026-09-30: 71.739 result · Public table; published model effort variant · evaluation date not published; verification date shown · source ↗
One observation; no trend can be inferred.
LiveBench code_generation / 2026-06-25 / %
- 2026-09-30: 73.239 result · Public table; published model effort variant · evaluation date not published; verification date shown · source ↗
One observation; no trend can be inferred.
LiveBench javascript / 2026-06-25 / %
- 2026-09-30: 72.727 result · Public table; published model effort variant · evaluation date not published; verification date shown · source ↗
One observation; no trend can be inferred.
LiveBench AMPS_Hard / 2026-06-25 / %
- 2026-09-30: 99 result · Public table; published model effort variant · evaluation date not published; verification date shown · source ↗
One observation; no trend can be inferred.
LiveBench olympiad / 2026-06-25 / %
- 2026-09-30: 92.159 result · Public table; published model effort variant · evaluation date not published; verification date shown · source ↗
One observation; no trend can be inferred.
LiveBench theory_of_mind / 2026-06-25 / %
- 2026-09-30: 76.923 result · Public table; published model effort variant · evaluation date not published; verification date shown · source ↗
One observation; no trend can be inferred.
LiveBench zebra_puzzle / 2026-06-25 / %
- 2026-09-30: 98.25 result · Public table; published model effort variant · evaluation date not published; verification date shown · source ↗
One observation; no trend can be inferred.
reasoning category score history
- 2026-09-30: 8 score · Scoring 3.0.0
One observation; no trend can be inferred.
agents category score history
- 2026-09-30: 0 score · Scoring 3.0.0
One observation; no trend can be inferred.
07 / EXPLORE