ZMIME
Comparison · 3 models · Updated Nov 22, 2026

GPT-6 Astra vs Gemini 3.1 Pro Preview vs Claude Fable 5.1

Too close to call on our weighted score (Gemini 3.1 Pro Preview 68, GPT-6 Astra 66, Claude Fable 5.1 65). The right pick depends on what you value most.Too close to call on our weighted score (GPT-6 Astra 89, Claude Fable 5.1 87, Gemini 3.1 Pro Preview 84). The right pick depends on what you value most.

  1. OpenAI

    GPT-6 Astra

    Released Sep 4, 2026

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

    Gemini 3.1 Pro Preview

    Released Feb 19, 2026Latest

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

    Claude Fable 5.1

    Released Sep 1, 2026

    65/100
    • ECI167.1
    • Price$10.00 / $50.00
    • Context1M
  4. OpenAI

    GPT-6 Astra

    Released Sep 4, 2026

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

    Gemini 3.1 Pro Preview

    Released Feb 19, 2026Latest

    84/100
    • ECI156.3
    • Price$2.00 / $12.00
    • Context1.05M
  6. Anthropic

    Claude Fable 5.1

    Released Sep 1, 2026

    87/100
    • ECI167.1
    • Price$10.00 / $50.00
    • Context1M
01 — Verdict

Too close to call

It is close. Our weighted score puts them within 1 points (Gemini 3.1 Pro Preview 68/100, GPT-6 Astra 66/100, Claude Fable 5.1 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%.It is close. Our weighted score puts them within 1 points (GPT-6 Astra 89/100, Claude Fable 5.1 87/100, Gemini 3.1 Pro Preview 84/100), 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 67%, inputs & features 20%, context window 13%.

  • CapabilityGPT-6 AstraCapabilities Index (ECI): GPT-6 Astra 169.2 · Claude Fable 5.1 167.1 · Gemini 3.1 Pro Preview 156.3
  • Lowest priceGemini 3.1 Pro PreviewGemini 3.1 Pro Preview $4.50 · GPT-6 Astra $20.00 · Claude Fable 5.1 $20.00 per 1M tokens (3:1 blend)
  • Longest contextGPT-6 Astra and Gemini 3.1 Pro PreviewGPT-6 Astra 1,050,000 · Gemini 3.1 Pro Preview 1,048,576 · Claude Fable 5.1 1,000,000 tokens
  • Widest inputsGemini 3.1 Pro PreviewGPT-6 Astra: Text, Images, PDFs · Gemini 3.1 Pro Preview: Text, Images, PDFs, Audio, Video · Claude Fable 5.1: Text, Images, PDFs
  • Self-hostingNo confirmed open weightsAll three are available only through APIs
How the score is built
MeasureWeightGPT-6 AstraGemini 3.1 Pro PreviewClaude Fable 5.1
CapabilityCapabilities Index (ECI)50%968494
Price25%0180
Inputs & features15%8010080
Context window10%656564
Overall100%66/10068/10065/100
How the score is built
MeasureWeightGPT-6 AstraGemini 3.1 Pro PreviewClaude Fable 5.1
CapabilityCapabilities Index (ECI)67%968494
Inputs & features20%8010080
Context window13%656564
Overall100%89/10084/10087/100
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-6 Astra vs Gemini 3.1 Pro Preview vs Claude Fable 5.1 specifications side by side
SpecificationGPT-6 AstraOpenAIGemini 3.1 Pro PreviewGoogleClaude Fable 5.1Anthropic
Capability
Capabilities Index (ECI)169.2 (best)156.3167.1
ECI rank#4 of 159 (best)#41 of 159#8 of 159
ECI confidence interval166.4–172.5154.5–158.2164.4–170.1
ECI evaluation dateSep 3, 2026Feb 19, 2026Sep 1, 2026
Documented capabilitiesGPT-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 solvingClaude model for demanding reasoning and long-horizon agentic work
Model parametersNot publishedNot publishedNot published
Output modalitiesTextTextText
GPQA DiamondGraduate-level science questions95.8%max (best)94.4%highNot published
FrontierMath Tiers 1–3Research-level mathematics93.7%max (best)59.7%90.2%max
OTIS Mock AIME 2024–2025Competition mathematics100%max (best)95.6%100%max (best)
SWE-bench VerifiedFixing real GitHub issuesNot published75.6%Not published
SimpleQA VerifiedShort factual questions75.6%max (best)73.5%high70.8%max
Publisher-reported benchmarks
Agents' Last ExamPublisher-reported result59.3Reported resultSep 3, 2026Not publishedNot published
OSWorld 2.0 · V2-Offline (v2026.08.08)Publisher-reported result72.6partial scoreReported resultSep 3, 2026Not publishedNot published
ScreenSpot-Prono tools92.7accuracyReported resultSep 3, 2026Not publishedNot published
AutomationBenchPublisher-reported result41.4success rateReported resultSep 3, 2026Not publishedNot published
BenchCADwith tools95.9geometric overlapReported resultSep 3, 2026Not publishedNot published
BrowseCompPublisher-reported result91.5accuracyReported resultSep 3, 2026Not publishedNot published
Terminal-Bench 4.0Publisher-reported result57.9success rateReported resultSep 3, 2026Not publishedNot published
DeepSWE 1.1Publisher-reported result74.1resolve rateReported resultSep 3, 2026Not publishedNot published
FrontierCode 1.1 · ExtendedPublisher-reported result64.5scoreReported resultSep 3, 2026Not publishedNot published
Terminal-Bench Science 0.1Publisher-reported result64.6success rateReported resultSep 3, 2026Not publishedNot published
FrontierMath v2 · Tier 4Publisher-reported result97.6accuracyReported resultSep 3, 2026Not publishedNot published
GPQA DiamondPublisher-reported result96accuracyReported resultSep 3, 202694.3accuracyReported resultApr 23, 2026Not published
Humanity's Last Examwith tools57.2accuracyReported resultSep 3, 2026Not publishedNot published
ExploitBenchwithout production safeguards100success rateReported resultSep 3, 2026Not publishedNot published
SRE-Benchsingle attempt88success rateReported resultSep 3, 2026Not publishedNot published
ARC-AGI-3Publisher-reported result99.9RHAEReported resultSep 3, 2026Not publishedNot published
Artificial Analysis Intelligence Index 4.1.1max61index scoreReported resultSep 3, 2026Not publishedNot published
Artificial Analysis Coding Agent Index 1.4max67index scoreReported resultSep 3, 2026Not publishedNot published
SWE-Bench ProPublisher-reported resultNot published54.2resolve rateReported resultMay 28, 2026Not published
Terminal-Bench 2.1Publisher-reported resultNot published70.3success rateReported resultMay 28, 2026Not published
SWE-Bench Pro · publicPublisher-reported resultNot published46.1resolve rateReported resultNot published
SWE-Atlas Codebase QnAPublisher-reported resultNot published13.5scoreReported resultNot published
SWE-Atlas RefactoringPublisher-reported resultNot published33.81scoreReported resultNot published
SWE-Atlas Test WritingPublisher-reported resultNot published29.84scoreReported resultNot published
Artificial Analysis Coding Agent IndexhighNot published43average pass@1Reported resultNot published
SWE-Atlas Codebase QnAhighNot published45.6pass@1Reported resultNot published
SWE-Bench Pro · hard-aahighNot published15.1pass@1Reported resultNot published
Terminal-Bench 2.1highNot published68.3pass@1Reported resultNot published
Humanity's Last Exam · full set, text + MMPublisher-reported resultNot published44.4accuracyReported resultMay 19, 2026Not published
ARC-AGI-2Publisher-reported resultNot published77.1accuracyReported resultMay 19, 2026Not published
MMMU Prono toolsNot published80.5accuracyReported resultMay 19, 2026Not published
MCP AtlasPublisher-reported resultNot published78.2success rateReported resultMay 19, 2026Not published
OSWorld-VerifiedPublisher-reported resultNot published76.2success rateReported resultMay 19, 2026Not published
CharXiv Reasoningno toolsNot published83.3accuracyReported resultMay 19, 2026Not published
GDPval-AAPublisher-reported resultNot published1314EloReported resultMay 19, 2026Not published
Terminal-Bench-Science 0.1production safeguards with fallbackNot publishedNot published52.6scoreReported resultSep 1, 2026
Terminal-Bench 4.0production safeguards with fallbackNot publishedNot published55.8scoreReported resultSep 1, 2026
GDPval-AA 2production safeguards with fallbackNot publishedNot published1853EloReported resultSep 1, 2026
OSWorld 2.0 · August 2026 task releaseproduction safeguards; interventions score zero; no fallbackNot publishedNot published77.9partial scoreReported resultSep 1, 2026
OSWorld 2.0 · August 2026 task releaseproduction safeguards; interventions score zero; no fallbackNot publishedNot published41.7strict success rateReported resultSep 1, 2026
Humanity's Last Examwithout tools; production safeguards with fallbackNot publishedNot published60.9scoreReported resultSep 1, 2026
Humanity's Last Examwith tools; production safeguards with fallbackNot publishedNot published65scoreReported resultSep 1, 2026
AutomationBenchproduction safeguards with fallbackNot publishedNot published31.4pass rateReported resultSep 1, 2026
CursorBench 3.2.0max effort; production safeguards with fallbackNot publishedNot published73.4scoreReported resultSep 1, 2026
Price per million tokens
Input$10.00$2.00 (best)$10.00
Output$50.00$12.00 (best)$50.00
Cached input$1.00$0.20 (best)$0.25
Blended (3:1)$20.00$4.50 (best)$20.00
Long-context rateOver 272K: $20.00 / $75.00Over 200K: $4.00 / $18.00Same rate
Price sourceOfficial OpenAI APIOfficial Google APIOfficial Anthropic API
Limits
Context window1,050,000 tokens (best)1,048,576 tokens1,000,000 tokens
Max input922,000 tokensNot publishedNot published
Max output128,000 tokens (best)65,536 tokens128,000 tokens (best)
Inputs and features
TextYesYesYes
ImagesYesYesYes
PDFsYesYesYes
AudioNoYesNo
VideoNoYesNo
ReasoningYeslow · medium · high · xhigh · maxYeslow · medium · highYeslow · medium · high · xhigh · max
Tool callingYesYesYes
Structured outputYesYesYes
File attachmentsYesYesYes
Temperature controlNoYesNo
Reasoning controleffortefforteffort
Availability
WeightsProprietaryProprietaryProprietary
API model IDgpt-6-astragemini-3.1-pro-previewclaude-fable-5-1
API providers56 (best)4051
ReleasedSep 4, 2026Feb 19, 2026Sep 1, 2026
Knowledge cutoffApr 30, 2026Jan 2025Jun 2026
Publisher documentationPublisher documentationPublisher documentationPublisher documentation
Source metadata updatedSep 4, 2026Feb 19, 2026Sep 1, 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-6 Astra$200.00
  • Gemini 3.1 Pro Preview$44.00
  • Claude Fable 5.1$200.00
04 — Questions

Which should you choose?

Which is better: GPT-6 Astra, Gemini 3.1 Pro Preview or Claude Fable 5.1?

It is close. Our weighted score puts them within 1 points (Gemini 3.1 Pro Preview 68/100, GPT-6 Astra 66/100, Claude Fable 5.1 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%.It is close. Our weighted score puts them within 1 points (GPT-6 Astra 89/100, Claude Fable 5.1 87/100, Gemini 3.1 Pro Preview 84/100), 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 67%, inputs & features 20%, context window 13%.

Which is cheaper, GPT-6 Astra, Gemini 3.1 Pro Preview or Claude Fable 5.1?

Gemini 3.1 Pro Preview is cheaper at $2.00 input / $12.00 output per million tokens (official Google API price). GPT-6 Astra costs $10.00 input / $50.00 output per million tokens (official OpenAI API price); Claude Fable 5.1 costs $10.00 input / $50.00 output per million tokens (official Anthropic 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 $20.00 for GPT-6 Astra (4.4× as much) and $20.00 for Claude Fable 5.1 (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), Claude Fable 5.1 167.1 (#8 of 159) and Gemini 3.1 Pro Preview 156.3 (#41 of 159). The confidence ranges of the top two overlap (166.4–172.5 vs 164.4–170.1), so treat the gap as small. On individual benchmarks: FrontierMath Tiers 1–3 — GPT-6 Astra 93.7%, Claude Fable 5.1 90.2%, Gemini 3.1 Pro Preview 59.7%; OTIS Mock AIME 2024–2025 — GPT-6 Astra 100%, Claude Fable 5.1 100%, Gemini 3.1 Pro Preview 95.6%; SimpleQA Verified — GPT-6 Astra 75.6%, Gemini 3.1 Pro Preview 73.5%, Claude Fable 5.1 70.8%.

Which is better for coding?

There are no published SWE-bench Verified results for GPT-6 Astra and Claude Fable 5.1 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?

GPT-6 Astra and Gemini 3.1 Pro Preview have the largest context windows (1,050,000 and 1,048,576 tokens), against 1,000,000 for Claude Fable 5.1. Maximum output per response: GPT-6 Astra up to 128,000, Gemini 3.1 Pro Preview up to 65,536, Claude Fable 5.1 up to 128,000 tokens.

Which can read images, PDFs, audio or video?

GPT-6 Astra accepts text, images and PDFs; Gemini 3.1 Pro Preview accepts text, images, PDFs, audio and video; Claude Fable 5.1 accepts text, images and PDFs. Gemini 3.1 Pro Preview handles the widest range of inputs.

Are any of these open source?

No. GPT-6 Astra, Gemini 3.1 Pro Preview and Claude Fable 5.1 are proprietary and only available through APIs and apps.

Which is newer?

GPT-6 Astra is the newest, released Sep 4, 2026. Claude Fable 5.1 came out Sep 1, 2026; Gemini 3.1 Pro Preview came out Feb 19, 2026. Knowledge cutoff: GPT-6 Astra Apr 30, 2026, Gemini 3.1 Pro Preview Jan 2025, Claude Fable 5.1 Jun 2026.

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.