Gemini 3.8 Flash vs GPT-5.4
Gemini 3.8 Flash leads GPT-5.4 by 4.6 points on the Shogo Agent Index, but their ranges overlap, so treat it as close. Gemini 3.8 Flash wins 13 and GPT-5.4 wins 5 of 19 shared benchmarks. Gemini 3.8 Flash is cheaper for agent workloads.
Add more models to this comparison →Pillars
| Pillar | Gemini 3.8 Flash | GPT-5.4 |
|---|---|---|
| Agentic coding | 63 | 57 |
| Long-horizon autonomy | 64 | 61 |
| Knowledge-work agents | 47 | 40 |
| Computer use | 44 | 42 |
| Security & context learning | 53 | 54 |
Shared benchmarks
| Benchmark | Gemini 3.8 Flash | GPT-5.4 |
|---|---|---|
| APEX-Agentsindex | 64.3% | 52.4% |
| DeepSWEindex | 73.8% | 51.8% |
| Vending-Bench 2index | $5,094 | $6,144 |
| ARC-AGI | 98.5% | 93.7% |
| ARC-AGI-2 | 89.2% | 74.0% |
| Chess Puzzles | 58.0% | 48.0% |
| CritPt | 18.3% | 23.4% |
| DTBench | 95.7% | 94.4% |
| FrontierMath Tier 4 | 22.0% | 49.0% |
| FrontierMath Tiers 1-3 | 68.4% | 78.6% |
| GPQA Diamond | 95.4% | 93.3% |
| Humanity's Last Exam | 44.5% | 36.2% |
| LMCA | 52.9% | 52.0% |
| MirrorCode | 38.9% | 23.3% |
| OTIS Mock AIME | 98.9% | 97.8% |
| ProofBench | 48.0% | 56.0% |
| SciCode | 56.6% | 56.6% |
| SimpleQA Verified | 69.7% | 45.1% |
| WebDev Arena | 1568 | 1463 |
Specs and price
| Gemini 3.8 Flash | GPT-5.4 | |
|---|---|---|
| Agent Index | 56.2 (#21) | 51.6 (#33) |
| Agent $/M | $0.35 | $1.27 |
| Input / output $/M | $0.75 / $3.75 | $2.50 / $15.00 |
| Context | 1M | 1.1M |
| Released | Sep 2, 2026 | Mar 5, 2026 |
| Weights | Closed | Closed |
| API hosts | 6 | 4 |
Don’t pick one. Route.
Shogo can run Gemini 3.8 Flash and GPT-5.4 side by side and send each step of an agent to the one that’s cheaper for it.
Try both in Shogo