Gemini 3.8 Flash vs GPT-6 Astra
AutomationBench-AA AutomationBench-AA: Agents orchestrate SaaS workflows through APIs; objectives receive credit only when the task has no guardrail violation
Gemini 3.8 Flash60.9
GPT-6 Astra64.6
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
Gemini 3.8 Flash46.1
GPT-6 Astra49.3
τ³-Banking · AA τ³-Banking · AA: Agents retrieve banking knowledge and execute tools to resolve customer-support workflows
Gemini 3.8 Flash45.8
GPT-6 Astra35.5
AA Analyst Agent AA Analyst Agent: Data-analyst tasks: the model explores data with tools and reports findings
Gemini 3.8 Flash · high46.3
GPT-6 Astra · max51.3
ITBench SRE ITBench SRE: IT incident tasks: diagnose and fix broken systems like a site-reliability engineer
Gemini 3.8 Flash · high52.5
GPT-6 Astra · max48.6
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 3.8 Flash19.7
GPT-6 Astra49.5
Terminal-Bench Science 0.1 Terminal-Bench Science 0.1: Agents complete expert-authored research workflows across five scientific domains in a terminal
Gemini 3.8 Flash · high10
GPT-6 Astra · max63.3
AA-Briefcase 1.1 · Elo AA-Briefcase 1.1 · Elo: Agents produce professional deliverables; blind pairwise comparisons produce the source Elo rating
Gemini 3.8 Flash1,181.5
GPT-6 Astra1,460.9
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)
Gemini 3.8 Flash84
GPT-6 Astra79.7
DeepSWE 1.1 DeepSWE 1.1: Agents change repository code to resolve software engineering tasks under the stated DeepSWE edition
Gemini 3.8 Flash71
GPT-6 Astra72.8
FrontierCode 1.1 FrontierCode 1.1: Agents implement changes in real repositories, graded against maintainer-defined requirements
Gemini 3.8 Flash · high36.4
GPT-6 Astra · max47
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
Gemini 3.8 Flash · high0
GPT-6 Astra · max8.6
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
Gemini 3.8 Flash · high94.9
GPT-6 Astra · high96.9
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
Gemini 3.8 Flash · high58.8
GPT-6 Astra · high90.2
Furniture Assembly Furniture Assembly: Models inspect furniture assembly photos, identify incorrect builds and locate the mistaken step
Gemini 3.8 Flash · high31.7
GPT-6 Astra · max80
AutomationBench · Zapier run AutomationBench · Zapier run: End-to-end completion of business workflows using real SaaS applications
Gemini 3.8 Flash—
GPT-6 Astra34.1
EBR Bench EBR Bench: Agents learn to play Earthborne Rangers across repeated playthroughs
Gemini 3.8 Flash—
GPT-6 Astra · max76.2
SpireClimb SpireClimb: Agents play Slay the Spire under the named game and difficulty setting
Gemini 3.8 Flash—
GPT-6 Astra · max25.7
SpireClimb StS1 A20 Heart SpireClimb StS1 A20 Heart: Agents play Slay the Spire under the named game and difficulty setting
Gemini 3.8 Flash—
GPT-6 Astra · max16.7
SpireClimb StS2 A10 SpireClimb StS2 A10: Agents play Slay the Spire under the named game and difficulty setting
Gemini 3.8 Flash—
GPT-6 Astra · max34.7
| Benchmark | Gemini 3.8 Flash | GPT-6 Astra |
|---|---|---|
| AutomationBench-AA | 60.9 | 64.6 |
| GDPval-AA (work tasks) 2.1 | 46.1 | 49.3 |
| τ³-Banking · AA | 45.8 | 35.5 |
| AA Analyst Agent | 46.3 · high | 51.3 · max |
| ITBench SRE | 52.5 · high | 48.6 · max |
| Terminal-Bench 4.0 | 19.7 | 49.5 |
| Terminal-Bench Science 0.1 | 10 · high | 63.3 · max |
| AA-Briefcase 1.1 · Elo | 1,181.5 | 1,460.9 |
| AA-LCR (long context) 1.1 | 84 | 79.7 |
| DeepSWE 1.1 | 71 | 72.8 |
| FrontierCode 1.1 | 36.4 · high | 47 · max |
| Harvey LAB-AA 1.1 | 0 · high | 8.6 · max |
| SWE-bench Pro (v2 full) | 94.9 · high | 96.9 · high |
| SWE-bench Pro (v2 hard) | 58.8 · high | 90.2 · high |
| Furniture Assembly | 31.7 · high | 80 · max |
| AutomationBench · Zapier run | — | 34.1 |
| EBR Bench | — | 76.2 · max |
| SpireClimb | — | 25.7 · max |
| SpireClimb StS1 A20 Heart | — | 16.7 · max |
| SpireClimb StS2 A10 | — | 34.7 · max |
Specs
| Spec | Gemini 3.8 Flash | GPT-6 Astra |
|---|---|---|
| Input, per 1M | !$0.75 | $10 |
| Output, per 1M | !$3.75 | $50 |
| Cost per task? | $0.93 | $1.54 |
| Speed, tok/s | — | 59 |
| Context | 1M | 1.05M |
| Weights | Closed | Closed |
| Released | Sep 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
Gemini 3.8 Flash → GPT-6 Astra
Monitoring can return HTTP 403›
Misalignment monitoring can return HTTP 403 before streaming or errors after output; match misalignment_policy_violation and handle stream errors.
No none reasoning effort›
The none reasoning effort is unsupported on GPT-6 Astra and GPT-6.1 Sol; use low instead.
Remove unsupported sampling parameters›
When reasoning effort is not none, remove temperature, top_p and top_logprobs, plus Chat Completions logprobs or Responses output logprobs inclusion.
GPT-6 Astra → Gemini 3.8 Flash
Context: 1.05M → 1M
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.