ZMIME
Comparison · 3 models · Updated Nov 22, 2026

GPT-5.4 vs GPT-6 Astra vs Gemini 3.1 Pro Preview

Too close to call on our weighted score (Gemini 3.1 Pro Preview 68, GPT-6 Astra 66, GPT-5.4 65). The right pick depends on what you value most.GPT-6 Astra comes out ahead, 89 to 84 and 82 on our weighted score, though Gemini 3.1 Pro Preview is 4.4× cheaper per token.

  1. OpenAI

    GPT-5.4

    Released Mar 5, 2026

    65/100
    • ECI157.9
    • Price$2.50 / $15.00
    • Context1.05M
  2. OpenAI

    GPT-6 Astra

    Released Sep 4, 2026

    66/100
    • ECI169.2
    • Price$10.00 / $50.00
    • Context1.05M
  3. Google

    Gemini 3.1 Pro Preview

    Released Feb 19, 2026Latest

    68/100
    • ECI156.3
    • Price$2.00 / $12.00
    • Context1.05M
  4. OpenAI

    GPT-5.4

    Released Mar 5, 2026

    82/100
    • ECI157.9
    • Price$2.50 / $15.00
    • Context1.05M
  5. Our pick

    OpenAI

    GPT-6 Astra

    Released Sep 4, 2026

    89/100
    • ECI169.2
    • Price$10.00 / $50.00
    • Context1.05M
  6. Google

    Gemini 3.1 Pro Preview

    Released Feb 19, 2026Latest

    84/100
    • ECI156.3
    • Price$2.00 / $12.00
    • Context1.05M
01 — Verdict

Too close to callGPT-6 Astra is our pick

It is close. Our weighted score puts them within 1 points (Gemini 3.1 Pro Preview 68/100, GPT-6 Astra 66/100, GPT-5.4 65/100), so choose by what matters most for your work: GPT-6 Astra for raw capability and Gemini 3.1 Pro Preview on price. The score weighs capability 50%, price 25%, inputs & features 15%, context window 10%.GPT-6 Astra is the better all-round choice, scoring 89/100 against Gemini 3.1 Pro Preview (84) and GPT-5.4 (82). It leads on capability. Gemini 3.1 Pro Preview wins on inputs & features. The score weighs capability 67%, inputs & features 20%, context window 13%.

  • CapabilityGPT-6 AstraCapabilities Index (ECI): GPT-6 Astra 169.2 · GPT-5.4 157.9 · Gemini 3.1 Pro Preview 156.3
  • Lowest priceGemini 3.1 Pro PreviewGemini 3.1 Pro Preview $4.50 · GPT-5.4 $5.63 · GPT-6 Astra $20.00 per 1M tokens (3:1 blend)
  • Longest contextAbout the sameGPT-5.4 1,050,000 · GPT-6 Astra 1,050,000 · Gemini 3.1 Pro Preview 1,048,576 tokens
  • Widest inputsGemini 3.1 Pro PreviewGPT-5.4: Text, Images, PDFs · GPT-6 Astra: Text, Images, PDFs · Gemini 3.1 Pro Preview: Text, Images, PDFs, Audio, Video
  • Self-hostingNo confirmed open weightsAll three are available only through APIs
How the score is built
MeasureWeightGPT-5.4GPT-6 AstraGemini 3.1 Pro Preview
CapabilityCapabilities Index (ECI)50%859684
Price25%14018
Inputs & features15%8080100
Context window10%656565
Overall100%65/10066/10068/100
How the score is built
MeasureWeightGPT-5.4GPT-6 AstraGemini 3.1 Pro Preview
CapabilityCapabilities Index (ECI)67%859684
Inputs & features20%8080100
Context window13%656565
Overall100%82/10089/10084/100
02 — Side by side

Every spec in one table

Highlighted cells lead their row. Unavailable specifications are marked “Not published”. Publisher results include their evaluation settings and source.

GPT-5.4 vs GPT-6 Astra vs Gemini 3.1 Pro Preview specifications side by side
SpecificationGPT-5.4OpenAIGPT-6 AstraOpenAIGemini 3.1 Pro PreviewGoogle
Capability
Capabilities Index (ECI)157.9169.2 (best)156.3
ECI rank#29 of 159#4 of 159 (best)#41 of 159
ECI confidence interval156.0–160.0166.4–172.5154.5–158.2
ECI evaluation dateMar 5, 2026Sep 3, 2026Feb 19, 2026
Documented capabilitiesAgent-ready GPT for coding and computer-use workflows at a lower costGPT-6 Astra is OpenAI's most capable model for complex reasoning, coding, computer use, research, and document creation.Reasoning-first Gemini preview for agentic coding and complex problem solving
Model parametersNot publishedNot publishedNot published
Output modalitiesTextTextText
GPQA DiamondGraduate-level science questions93.3%xhigh95.8%max (best)94.4%high
FrontierMath Tiers 1–3Research-level mathematics78.6%xhigh93.7%max (best)59.7%
OTIS Mock AIME 2024–2025Competition mathematics97.8%high100%max (best)95.6%
SWE-bench VerifiedFixing real GitHub issues76.9%high (best)Not published75.6%
SimpleQA VerifiedShort factual questions45.1%xhigh75.6%max (best)73.5%high
Publisher-reported benchmarks
SWE-Bench Pro · publicPublisher-reported result59.1resolve rateReported resultNot published46.1resolve rateReported result
SWE-Atlas Codebase QnAPublisher-reported result40.8scoreReported resultNot published13.5scoreReported result
SWE-Atlas RefactoringPublisher-reported result44.29scoreReported resultNot published33.81scoreReported result
SWE-Atlas Test WritingPublisher-reported result44.36scoreReported resultNot published29.84scoreReported result
Artificial Analysis Coding Agent Indexmedium53.6average pass@1Reported resultNot publishedNot published
SWE-Atlas Codebase QnAmedium72.4pass@1Reported resultNot publishedNot published
SWE-Bench Pro · hard-aamedium18.4pass@1Reported resultNot publishedNot published
Terminal-Bench 2.1medium69.8pass@1Reported resultNot publishedNot published
Terminal-Bench 2.0Publisher-reported result75.1success rateReported resultApr 23, 2026Not publishedNot published
GPQA DiamondPublisher-reported result92.8accuracyReported resultApr 23, 202696accuracyReported resultSep 3, 202694.3accuracyReported resultApr 23, 2026
Humanity's Last Examno tools39.8accuracyReported resultApr 23, 2026Not publishedNot published
Humanity's Last Examwith tools52.1accuracyReported resultApr 23, 202657.2accuracyReported resultSep 3, 2026Not published
OSWorld-VerifiedPublisher-reported result75success rateReported resultApr 23, 2026Not published76.2success rateReported resultMay 19, 2026
BrowseCompPublisher-reported result82.7accuracyReported resultApr 23, 202691.5accuracyReported resultSep 3, 2026Not published
GDPvalPublisher-reported result83wins or tiesReported resultApr 23, 2026Not publishedNot published
ARC-AGI-2Verified73.3accuracyReported resultApr 23, 2026Not publishedNot published
FrontierMath · Tier 1-3Publisher-reported result47.6accuracyReported resultApr 23, 2026Not publishedNot published
FrontierMath · Tier 4Publisher-reported result27.1accuracyReported resultApr 23, 2026Not publishedNot published
MMMU Prono tools81.2accuracyReported resultApr 23, 2026Not published80.5accuracyReported resultMay 19, 2026
Agents' Last ExamPublisher-reported resultNot published59.3Reported resultSep 3, 2026Not published
OSWorld 2.0 · V2-Offline (v2026.08.08)Publisher-reported resultNot published72.6partial scoreReported resultSep 3, 2026Not published
ScreenSpot-Prono toolsNot published92.7accuracyReported resultSep 3, 2026Not published
AutomationBenchPublisher-reported resultNot published41.4success rateReported resultSep 3, 2026Not published
BenchCADwith toolsNot published95.9geometric overlapReported resultSep 3, 2026Not published
Terminal-Bench 4.0Publisher-reported resultNot published57.9success rateReported resultSep 3, 2026Not published
DeepSWE 1.1Publisher-reported resultNot published74.1resolve rateReported resultSep 3, 2026Not published
FrontierCode 1.1 · ExtendedPublisher-reported resultNot published64.5scoreReported resultSep 3, 2026Not published
Terminal-Bench Science 0.1Publisher-reported resultNot published64.6success rateReported resultSep 3, 2026Not published
FrontierMath v2 · Tier 4Publisher-reported resultNot published97.6accuracyReported resultSep 3, 2026Not published
ExploitBenchwithout production safeguardsNot published100success rateReported resultSep 3, 2026Not published
SRE-Benchsingle attemptNot published88success rateReported resultSep 3, 2026Not published
ARC-AGI-3Publisher-reported resultNot published99.9RHAEReported resultSep 3, 2026Not published
Artificial Analysis Intelligence Index 4.1.1maxNot published61index scoreReported resultSep 3, 2026Not published
Artificial Analysis Coding Agent Index 1.4maxNot published67index scoreReported resultSep 3, 2026Not published
SWE-Bench ProPublisher-reported resultNot publishedNot published54.2resolve rateReported resultMay 28, 2026
Terminal-Bench 2.1Publisher-reported resultNot publishedNot published70.3success rateReported resultMay 28, 2026
Artificial Analysis Coding Agent IndexhighNot publishedNot published43average pass@1Reported result
SWE-Atlas Codebase QnAhighNot publishedNot published45.6pass@1Reported result
SWE-Bench Pro · hard-aahighNot publishedNot published15.1pass@1Reported result
Terminal-Bench 2.1highNot publishedNot published68.3pass@1Reported result
Humanity's Last Exam · full set, text + MMPublisher-reported resultNot publishedNot published44.4accuracyReported resultMay 19, 2026
ARC-AGI-2Publisher-reported resultNot publishedNot published77.1accuracyReported resultMay 19, 2026
MCP AtlasPublisher-reported resultNot publishedNot published78.2success rateReported resultMay 19, 2026
CharXiv Reasoningno toolsNot publishedNot published83.3accuracyReported resultMay 19, 2026
GDPval-AAPublisher-reported resultNot publishedNot published1314EloReported resultMay 19, 2026
Price per million tokens
Input$2.50$10.00$2.00 (best)
Output$15.00$50.00$12.00 (best)
Cached input$0.25$1.00$0.20 (best)
Blended (3:1)$5.63$20.00$4.50 (best)
Long-context rateOver 272K: $5.00 / $22.50Over 272K: $20.00 / $75.00Over 200K: $4.00 / $18.00
Price sourceOfficial OpenAI APIOfficial OpenAI APIOfficial Google API
Limits
Context window1,050,000 tokens (best)1,050,000 tokens (best)1,048,576 tokens
Max input922,000 tokens922,000 tokensNot published
Max output128,000 tokens (best)128,000 tokens (best)65,536 tokens
Inputs and features
TextYesYesYes
ImagesYesYesYes
PDFsYesYesYes
AudioNoNoYes
VideoNoNoYes
ReasoningYeslow · medium · high · xhighYeslow · medium · high · xhigh · maxYeslow · medium · high
Tool callingYesYesYes
Structured outputYesYesYes
File attachmentsYesYesYes
Temperature controlYesNoYes
Reasoning controleffortefforteffort
Availability
WeightsProprietaryProprietaryProprietary
API model IDgpt-5.4gpt-6-astragemini-3.1-pro-preview
API providers5056 (best)40
ReleasedMar 5, 2026Sep 4, 2026Feb 19, 2026
Knowledge cutoffAug 31, 2025Apr 30, 2026Jan 2025
Publisher documentationPublisher documentationPublisher documentationPublisher documentation
Source metadata updatedMar 5, 2026Sep 4, 2026Feb 19, 2026
Data sourcesmodels.dev · Epoch AImodels.dev · Epoch AImodels.dev · Epoch AI
Catalog checkedNov 22, 2026Nov 22, 2026Nov 22, 2026
03 — Cost

What would a month cost?

Enter your expected volume in millions of tokens. List prices only; caching and batch discounts would lower these.

  • GPT-5.4$55.00
  • GPT-6 Astra$200.00
  • Gemini 3.1 Pro Preview$44.00
04 — Questions

Which should you choose?

Which is better: GPT-5.4, GPT-6 Astra or Gemini 3.1 Pro Preview?

It is close. Our weighted score puts them within 1 points (Gemini 3.1 Pro Preview 68/100, GPT-6 Astra 66/100, GPT-5.4 65/100), so choose by what matters most for your work: GPT-6 Astra for raw capability and Gemini 3.1 Pro Preview on price. The score weighs capability 50%, price 25%, inputs & features 15%, context window 10%.GPT-6 Astra is the better all-round choice, scoring 89/100 against Gemini 3.1 Pro Preview (84) and GPT-5.4 (82). It leads on capability. Gemini 3.1 Pro Preview wins on inputs & features. The score weighs capability 67%, inputs & features 20%, context window 13%.

Which is cheaper, GPT-5.4, GPT-6 Astra or Gemini 3.1 Pro Preview?

Gemini 3.1 Pro Preview is cheaper at $2.00 input / $12.00 output per million tokens (official Google API price). GPT-5.4 costs $2.50 input / $15.00 output per million tokens (official OpenAI API price); GPT-6 Astra costs $10.00 input / $50.00 output per million tokens (official OpenAI API price). At a typical mix of three input tokens to one output token, that is $4.50 per million tokens for Gemini 3.1 Pro Preview versus $5.63 for GPT-5.4 (1.3× as much) and $20.00 for GPT-6 Astra (4.4× as much).

Which scores higher on benchmarks?

GPT-6 Astra scores higher on the Capabilities Index (ECI): GPT-6 Astra 169.2 (#4 of 159), GPT-5.4 157.9 (#29 of 159) and Gemini 3.1 Pro Preview 156.3 (#41 of 159). Their confidence ranges do not overlap (166.4–172.5 vs 156.0–160.0), so the gap is a real one. On individual benchmarks: GPQA Diamond — GPT-6 Astra 95.8%, Gemini 3.1 Pro Preview 94.4%, GPT-5.4 93.3%; FrontierMath Tiers 1–3 — GPT-6 Astra 93.7%, GPT-5.4 78.6%, Gemini 3.1 Pro Preview 59.7%; OTIS Mock AIME 2024–2025 — GPT-6 Astra 100%, GPT-5.4 97.8%, Gemini 3.1 Pro Preview 95.6%; SimpleQA Verified — GPT-6 Astra 75.6%, Gemini 3.1 Pro Preview 73.5%, GPT-5.4 45.1%.

Which is better for coding?

There are no published SWE-bench Verified results for GPT-6 Astra yet, so there is no like-for-like coding score. On overall capability, GPT-6 Astra leads, which tends to carry over to coding, but test on your own codebase. All three support tool calling for agent workflows.

Which has the bigger context window?

Their context windows are effectively the same size: GPT-5.4 1,050,000, GPT-6 Astra 1,050,000 and Gemini 3.1 Pro Preview 1,048,576 tokens. Maximum output per response: GPT-5.4 up to 128,000, GPT-6 Astra up to 128,000, Gemini 3.1 Pro Preview up to 65,536 tokens.

Which can read images, PDFs, audio or video?

GPT-5.4 accepts text, images and PDFs; GPT-6 Astra accepts text, images and PDFs; Gemini 3.1 Pro Preview accepts text, images, PDFs, audio and video. Gemini 3.1 Pro Preview handles the widest range of inputs.

Are any of these open source?

No. GPT-5.4, GPT-6 Astra and Gemini 3.1 Pro Preview are proprietary and only available through APIs and apps.

Which is newer?

GPT-6 Astra is the newest, released Sep 4, 2026. GPT-5.4 came out Mar 5, 2026; Gemini 3.1 Pro Preview came out Feb 19, 2026. Knowledge cutoff: GPT-5.4 Aug 31, 2025, GPT-6 Astra Apr 30, 2026, Gemini 3.1 Pro Preview Jan 2025.

How do you decide the winner?

This rating uses capability (50%, independent benchmark results); price (25%, blended price per million tokens (3 input : 1 output), log scale); inputs & features (15%, image, PDF, audio and video input, tool calling, structured output and reasoning); context window (10%, maximum tokens per request, log scale). Dimensions missing for any model are dropped and the remaining weights rescaled, so every model is judged on the same evidence. Specs and prices come from public model listings and the labs’ own API pages; capability scores come from independent benchmark runs. Data updated Nov 22, 2026.This rating uses capability (67%, independent benchmark results); inputs & features (20%, image, PDF, audio and video input, tool calling, structured output and reasoning); context window (13%, maximum tokens per request, log scale). Dimensions missing for any model are dropped and the remaining weights rescaled, so every model is judged on the same evidence. Specs and prices come from public model listings and the labs’ own API pages; capability scores come from independent benchmark runs. Data updated Nov 22, 2026.