Gemini 3.6 Flash vs Gemini 3.8 Flash
Different efforts: high · medium · Compare all at high
Humanity's Last Exam Humanity's Last Exam: Very hard expert questions across many fields, written to stump today’s models
Gemini 3.6 Flash40.8
Gemini 3.8 Flash42.1
Terminal-Bench 4.0 Terminal-Bench 4.0 · source setting 2: Agents use a terminal to complete executable coding, system administration and research tasks in the 4.0 suite
Gemini 3.6 Flash7.1
Gemini 3.8 Flash19.7
AutomationBench-AA AutomationBench-AA: Agents orchestrate SaaS workflows through APIs; objectives receive credit only when the task has no guardrail violation
Gemini 3.6 Flash53
Gemini 3.8 Flash60.9
Arena Text · creative writing Arena Text · creative writing: Human pairwise preferences in the stated text vote pool
Gemini 3.6 Flash1,471
Gemini 3.8 Flash · high1,487
MMMU-Pro (vision) MMMU-Pro (vision): College-level questions that require understanding images: charts, diagrams, photos
Gemini 3.6 Flash83.2
Gemini 3.8 Flash84.2
| Benchmark | Gemini 3.6 Flash | Gemini 3.8 Flash |
|---|---|---|
| Humanity's Last Exam | 40.8 | 42.1 |
| Terminal-Bench 4.0 | 7.1 | 19.7 |
| AutomationBench-AA | 53 | 60.9 |
| Arena Text · creative writing | 1,471 | 1,487 · high |
| MMMU-Pro (vision) | 83.2 | 84.2 |
Specs
| Spec | Gemini 3.6 Flash | Gemini 3.8 Flash |
|---|---|---|
| Input, per 1M | !$0.75 | !$0.75 |
| Output, per 1M | !$3.75 | !$3.75 |
| Cost per task? | $1.60 | $0.93 |
| Speed, tok/s | 209 | — |
| Context | 1M | 1M |
| Weights | Closed | Closed |
| Released | Jul 2026 | Sep 2026 |
| See images | ✓ | ✓ |
| Read PDFs | ✓ | ✓ |
| Hear audio | ✓ | ✓ |
| Watch video | ✓ | ✓ |
| Call tools | ✓ | ✓ |
| Force a tool call | Partial | Partial |
| Return strict JSON | ✓ | ✓ |
| Search the web | ✓ | ✓ |
| Run code | ✓ | ✓ |
| Use a computer | Partial | ? |
What breaks when you switch
Gemini 3.6 Flash → Gemini 3.8 Flash
Sampling parameters are ignored›
temperature, top_p and top_k are ignored now and will cause HTTP 400 in future generations; remove them from requests.
No minimal thinking level›
The minimal thinking level returns an error; avoid minimal.
Gemini 3.8 Flash → Gemini 3.6 Flash
Nothing known.