GPT-6 Astra vs MiniMax M3.1 Flash
Different efforts: medium · thinking
AutomationBench-AA AutomationBench-AA: Agents orchestrate SaaS workflows through APIs; objectives receive credit only when the task has no guardrail violation
GPT-6 Astra64.6
MiniMax M3.1 Flash72.5
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
GPT-6 Astra49.3
MiniMax M3.1 Flash47.9
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
GPT-6 Astra49.5
MiniMax M3.1 Flash25.8
AA-Briefcase 1.1 · Elo AA-Briefcase 1.1 · Elo: Agents produce professional deliverables; blind pairwise comparisons produce the source Elo rating
GPT-6 Astra1,460.9
MiniMax M3.1 Flash1,473.9
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)
GPT-6 Astra79.7
MiniMax M3.1 Flash79.7
FrontierCode 1.1 FrontierCode 1.1: Agents implement changes in real repositories, graded against maintainer-defined requirements
GPT-6 Astra · max47
MiniMax M3.1 Flash38.6
AutomationBench · Zapier run AutomationBench · Zapier run: End-to-end completion of business workflows using real SaaS applications
GPT-6 Astra34.1
MiniMax M3.1 Flash—
τ³-Banking · AA τ³-Banking · AA: Agents retrieve banking knowledge and execute tools to resolve customer-support workflows
GPT-6 Astra35.5
MiniMax M3.1 Flash—
AA Analyst Agent AA Analyst Agent: Data-analyst tasks: the model explores data with tools and reports findings
GPT-6 Astra · max51.3
MiniMax M3.1 Flash—
ITBench SRE ITBench SRE: IT incident tasks: diagnose and fix broken systems like a site-reliability engineer
GPT-6 Astra · max48.6
MiniMax M3.1 Flash—
Terminal-Bench Science 0.1 Terminal-Bench Science 0.1: Agents complete expert-authored research workflows across five scientific domains in a terminal
GPT-6 Astra · max63.3
MiniMax M3.1 Flash—
DeepSWE 1.1 DeepSWE 1.1: Agents change repository code to resolve software engineering tasks under the stated DeepSWE edition
GPT-6 Astra72.8
MiniMax M3.1 Flash—
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
GPT-6 Astra · max8.6
MiniMax M3.1 Flash—
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
GPT-6 Astra · high96.9
MiniMax M3.1 Flash—
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
GPT-6 Astra · high90.2
MiniMax M3.1 Flash—
EBR Bench EBR Bench: Agents learn to play Earthborne Rangers across repeated playthroughs
GPT-6 Astra · max76.2
MiniMax M3.1 Flash—
Furniture Assembly Furniture Assembly: Models inspect furniture assembly photos, identify incorrect builds and locate the mistaken step
GPT-6 Astra · max80
MiniMax M3.1 Flash—
SpireClimb SpireClimb: Agents play Slay the Spire under the named game and difficulty setting
GPT-6 Astra · max25.7
MiniMax M3.1 Flash—
SpireClimb StS1 A20 Heart SpireClimb StS1 A20 Heart: Agents play Slay the Spire under the named game and difficulty setting
GPT-6 Astra · max16.7
MiniMax M3.1 Flash—
SpireClimb StS2 A10 SpireClimb StS2 A10: Agents play Slay the Spire under the named game and difficulty setting
GPT-6 Astra · max34.7
MiniMax M3.1 Flash—
| Benchmark | GPT-6 Astra | MiniMax M3.1 Flash |
|---|---|---|
| AutomationBench-AA | 64.6 | 72.5 |
| GDPval-AA (work tasks) 2.1 | 49.3 | 47.9 |
| Terminal-Bench 4.0 | 49.5 | 25.8 |
| AA-Briefcase 1.1 · Elo | 1,460.9 | 1,473.9 |
| AA-LCR (long context) 1.1 | 79.7 | 79.7 |
| FrontierCode 1.1 | 47 · max | 38.6 |
| AutomationBench · Zapier run | 34.1 | — |
| τ³-Banking · AA | 35.5 | — |
| AA Analyst Agent | 51.3 · max | — |
| ITBench SRE | 48.6 · max | — |
| Terminal-Bench Science 0.1 | 63.3 · max | — |
| DeepSWE 1.1 | 72.8 | — |
| Harvey LAB-AA 1.1 | 8.6 · max | — |
| SWE-bench Pro (v2 full) | 96.9 · high | — |
| SWE-bench Pro (v2 hard) | 90.2 · high | — |
| EBR Bench | 76.2 · max | — |
| Furniture Assembly | 80 · max | — |
| SpireClimb | 25.7 · max | — |
| SpireClimb StS1 A20 Heart | 16.7 · max | — |
| SpireClimb StS2 A10 | 34.7 · max | — |
Specs
| Spec | GPT-6 Astra | MiniMax M3.1 Flash |
|---|---|---|
| Input, per 1M | $10 | $0.15 |
| Output, per 1M | $50 | $0.60 |
| Cost per task? | $1.54 | $0.22 |
| Speed, tok/s | 59 | 105 |
| Context | 1.05M | 1M |
| Weights | Closed | Closed |
| Released | Sep 2026 | Oct 2026 |
| See images | ✓ | ✓ |
| Read PDFs | Partial | — |
| Hear audio | — | — |
| Watch video | — | ✓ |
| Call tools | ✓ | Partial |
| Force a tool call | Partial | ? |
| Return strict JSON | ✓ | Partial |
| Search the web | ✓ | ? |
| Run code | ✓ | ? |
| Use a computer | ✓ | ? |
What breaks when you switch
GPT-6 Astra → MiniMax M3.1 Flash
Context: 1.05M → 1M
Complete model response must be appended›
In multi-turn tools, preserve the complete assistant response, including reasoning_content and tool_calls, to maintain reasoning continuity.
MiniMax M3.1 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.