Claude Opus 5 vs Gemini 3.8 Flash
AutomationBench-AA AutomationBench-AA: Agents orchestrate SaaS workflows through APIs; objectives receive credit only when the task has no guardrail violation
Opus 554.3
Gemini 3.8 Flash60.9
GDPval-AA (work tasks) 2.1 GDPval-AA (work tasks) 2.1: Real work tasks from many professions (reports, spreadsheets, analyses), graded against expert work
Opus 549.6
Gemini 3.8 Flash46.1
τ³-Banking · AA τ³-Banking · AA: Agents retrieve banking knowledge and execute tools to resolve customer-support workflows
Opus 538.6
Gemini 3.8 Flash45.8
AA Analyst Agent AA Analyst Agent: Data-analyst tasks: the model explores data with tools and reports findings
Opus 5 · max53.8
Gemini 3.8 Flash · high46.3
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
Opus 534.3
Gemini 3.8 Flash19.7
Terminal-Bench Science 0.1 Terminal-Bench Science 0.1: Agents complete expert-authored research workflows across five scientific domains in a terminal
Opus 5 · max28.6
Gemini 3.8 Flash · high10
AA-Briefcase 1.1 · Elo AA-Briefcase 1.1 · Elo: Agents produce professional deliverables; blind pairwise comparisons produce the source Elo rating
Opus 51,434.9
Gemini 3.8 Flash1,181.5
AA-LCR (long context) 1.1 AA-LCR (long context) 1.1: Questions that need reading and connecting facts across very long documents (around 100K tokens)
Opus 582
Gemini 3.8 Flash84
DeepSWE 1.1 DeepSWE 1.1: Agents change repository code to resolve software engineering tasks under the stated DeepSWE edition
Opus 568.9
Gemini 3.8 Flash71
SWE-bench Pro (v2 full) SWE-bench Pro (v2 full): Agents resolve repository issues under the network-locked v2 protocol; full and 51-task Hard results are separate
Opus 5 · xhigh99.4
Gemini 3.8 Flash · high94.9
SWE-bench Pro (v2 hard) SWE-bench Pro (v2 hard): Agents resolve repository issues under the network-locked v2 protocol; full and 51-task Hard results are separate
Opus 5 · xhigh98
Gemini 3.8 Flash · high58.8
Furniture Assembly Furniture Assembly: Models inspect furniture assembly photos, identify incorrect builds and locate the mistaken step
Opus 5 · max60.8
Gemini 3.8 Flash · high31.7
τ³-bench · banking · 1.0.1 τ-bench official · 3-banking-1.0.1: Customer-service agents converse with simulated users, retrieve domain knowledge and change state through tools
Opus 5 · max48.7
Gemini 3.8 Flash—
ITBench SRE ITBench SRE: IT incident tasks: diagnose and fix broken systems like a site-reliability engineer
Opus 5—
Gemini 3.8 Flash · high52.5
EnterpriseOps-Gym-AA EnterpriseOps-Gym-AA: Agents complete stateful enterprise workflows in eight domains; final database state determines success
Opus 5 · max47.5
Gemini 3.8 Flash—
FrontierCode 1.1 FrontierCode 1.1: Agents implement changes in real repositories, graded against maintainer-defined requirements
Opus 5—
Gemini 3.8 Flash · high36.4
Harvey LAB-AA 1.1 Harvey LAB-AA 1.1: Legal agents work from case documents; every rubric criterion must pass and material hallucinations receive no credit
Opus 5—
Gemini 3.8 Flash · high0
EBR Bench EBR Bench: Agents learn to play Earthborne Rangers across repeated playthroughs
Opus 5 · max45.7
Gemini 3.8 Flash—
| Benchmark | Opus 5 | Gemini 3.8 Flash |
|---|---|---|
| AutomationBench-AA | 54.3 | 60.9 |
| GDPval-AA (work tasks) 2.1 | 49.6 | 46.1 |
| τ³-Banking · AA | 38.6 | 45.8 |
| AA Analyst Agent | 53.8 · max | 46.3 · high |
| Terminal-Bench 4.0 | 34.3 | 19.7 |
| Terminal-Bench Science 0.1 | 28.6 · max | 10 · high |
| AA-Briefcase 1.1 · Elo | 1,434.9 | 1,181.5 |
| AA-LCR (long context) 1.1 | 82 | 84 |
| DeepSWE 1.1 | 68.9 | 71 |
| SWE-bench Pro (v2 full) | 99.4 · xhigh | 94.9 · high |
| SWE-bench Pro (v2 hard) | 98 · xhigh | 58.8 · high |
| Furniture Assembly | 60.8 · max | 31.7 · high |
| τ³-bench · banking · 1.0.1 | 48.7 · max | — |
| ITBench SRE | — | 52.5 · high |
| EnterpriseOps-Gym-AA | 47.5 · max | — |
| FrontierCode 1.1 | — | 36.4 · high |
| Harvey LAB-AA 1.1 | — | 0 · high |
| EBR Bench | 45.7 · max | — |
Specs
| Spec | Opus 5 | Gemini 3.8 Flash |
|---|---|---|
| Input, per 1M | $5 | !$0.75 |
| Output, per 1M | $25 | !$3.75 |
| Cost per task? | $2.19 | $0.93 |
| Speed, tok/s | 55 | — |
| Context | 1M | 1M |
| Weights | Closed | Closed |
| Released | Jul 2026 | Sep 2026 |
| See images | ✓ | ✓ |
| Read PDFs | Partial | ✓ |
| Hear audio | — | ✓ |
| Watch video | — | ✓ |
| Call tools | ✓ | ✓ |
| Force a tool call | Partial | Partial |
| Return strict JSON | ✓ | ✓ |
| Search the web | ? | ✓ |
| Run code | ? | ✓ |
| Use a computer | ? | ? |
What breaks when you switch
Opus 5 → 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 → Opus 5
Nothing known.