ZMIME
Comparison · 3 models · Updated Nov 18, 2026

GPT-6.2 Astra vs Step 5 Preview vs Claude Fable 5.2

No independent capability benchmark covers every model here yet. The score uses price, inputs & features and context window.

Price in rating: Included · Exclude price

  1. OpenAI

    GPT-6.2 Astra

    Released Nov 17, 2026New

    37/100
    • ECI174.2
    • Price$8.00 / $40.00
    • Context1.05M
  2. StepFun

    Step 5 Preview

    Released Sep 16, 2026

    57/100
    • ECI—
    • Price$1.00 / $2.70
    • Context1M
  3. Anthropic

    Claude Fable 5.2

    Released Nov 9, 2026New

    37/100
    • ECI173.5
    • Price$8.00 / $40.00
    • Context1M
01 — Verdict

Capability evidence is incomplete

There is not enough shared capability evidence to name an overall winner. The scores (Step 5 Preview 57/100, GPT-6.2 Astra 37/100, Claude Fable 5.2 37/100) compare the available specifications and price. The score weighs price 50%, inputs & features 30%, context window 20%. None of the shared benchmarks cover every model here yet, so capability is left out of this verdict.

  • CapabilityNot enough dataNo independent benchmark covers every model here yet
  • Lowest priceStep 5 PreviewStep 5 Preview $1.43 · GPT-6.2 Astra $16.00 · Claude Fable 5.2 $16.00 per 1M tokens (3:1 blend)
  • Longest contextGPT-6.2 AstraGPT-6.2 Astra 1,050,000 · Step 5 Preview 1,000,000 · Claude Fable 5.2 1,000,000 tokens
  • Widest inputsSame inputsGPT-6.2 Astra: Text, Images, PDFs · Step 5 Preview: Text, Images, Video · Claude Fable 5.2: Text, Images, PDFs
  • Self-hostingStep 5 PreviewPublishes downloadable weights
How the score is built
MeasureWeightGPT-6.2 AstraStep 5 PreviewClaude Fable 5.2
Price50%0410
Inputs & features30%808080
Context window20%656464
Specs & price only100%37/10057/10037/100

Left out because at least one model lacks the data: capability. The remaining weights were rescaled.

02 — Side by side

Every spec in one table

Highlighted cells lead their row. Dashes mean the data is not published.

GPT-6.2 Astra vs Step 5 Preview vs Claude Fable 5.2 specifications side by side
SpecificationGPT-6.2 AstraOpenAIStep 5 PreviewStepFunClaude Fable 5.2Anthropic
Capability
Capabilities Index (ECI)174.2 (best)—173.5
ECI rank#1 of 158 (best)—#2 of 158
GPQA DiamondGraduate-level science questions96.8% (best)—96.0%
FrontierMath Tiers 1–3Research-level mathematics92.6%—95.1% (best)
OTIS Mock AIME 2024–2025Competition mathematics100% (best)—99.7%
SimpleQA VerifiedShort factual questions78.5%—94.0% (best)
Price per million tokens
Input$8.00$1.00 (best)$8.00
Output$40.00$2.70 (best)$40.00
Cached input$0.40$0.10 (best)$0.20
Blended (3:1)$16.00$1.43 (best)$16.00
Long-context rateOver 272K: $16.00 / $60.00Same rateSame rate
Price sourceOfficial OpenAI APIOfficial StepFun (Global) APIOfficial Anthropic API
Limits
Context window1,050,000 tokens (best)1,000,000 tokens1,000,000 tokens
Max output128,000 tokens1,000,000 tokens (best)128,000 tokens
Inputs and features
TextYesYesYes
ImagesYesYesYes
PDFsYesNoYes
AudioNoNoNo
VideoNoYesNo
ReasoningYeslow · medium · high · xhigh · maxYeslow · medium · highYeslow · medium · high · xhigh · max
Tool callingYesYesYes
Structured outputYesYesYes
Availability
WeightsProprietaryOpenProprietary
API model IDgpt-6.2-astrastep-5-previewclaude-fable-5-2
API providers122125 (best)
ReleasedNov 17, 2026Sep 16, 2026Nov 9, 2026
Knowledge cutoffMay 18, 2026—Jun 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-6.2 Astra$160.00
  • Step 5 Preview$15.40
  • Claude Fable 5.2$160.00
04 — Questions

Which should you choose?

Which is better: GPT-6.2 Astra, Step 5 Preview or Claude Fable 5.2?

There is not enough shared capability evidence to name an overall winner. The scores (Step 5 Preview 57/100, GPT-6.2 Astra 37/100, Claude Fable 5.2 37/100) compare the available specifications and price. The score weighs price 50%, inputs & features 30%, context window 20%. None of the shared benchmarks cover every model here yet, so capability is left out of this verdict.

Which is cheaper, GPT-6.2 Astra, Step 5 Preview or Claude Fable 5.2?

Step 5 Preview is cheaper at $1.00 input / $2.70 output per million tokens (official StepFun (Global) API price). GPT-6.2 Astra costs $8.00 input / $40.00 output per million tokens (official OpenAI API price); Claude Fable 5.2 costs $8.00 input / $40.00 output per million tokens (official Anthropic API price). At a typical mix of three input tokens to one output token, that is $1.43 per million tokens for Step 5 Preview versus $16.00 for GPT-6.2 Astra (11× as much) and $16.00 for Claude Fable 5.2 (11× as much).

Which scores higher on benchmarks?

There is no independent benchmark that covers all three models yet. GPT-6.2 Astra has an ECI of 174.2, Step 5 Preview has not been scored yet and Claude Fable 5.2 has an ECI of 173.5.

Which is better for coding?

There are no published SWE-bench Verified results for GPT-6.2 Astra, Step 5 Preview and Claude Fable 5.2 yet, so there is no like-for-like coding score. Test both on a sample of your own tasks. All three support tool calling for agent workflows.

Which has the bigger context window?

GPT-6.2 Astra has the largest context window at 1,050,000 tokens, against 1,000,000 for Step 5 Preview and 1,000,000 for Claude Fable 5.2. Maximum output per response: GPT-6.2 Astra up to 128,000, Step 5 Preview up to 1,000,000, Claude Fable 5.2 up to 128,000 tokens.

Which can read images, PDFs, audio or video?

GPT-6.2 Astra accepts text, images and PDFs; Step 5 Preview accepts text, images and video; Claude Fable 5.2 accepts text, images and PDFs. They handle the same number of input types.

Are any of these open source?

Step 5 Preview publishes its weights and can be self-hosted; GPT-6.2 Astra and Claude Fable 5.2 is proprietary.

Which is newer?

GPT-6.2 Astra is the newest, released Nov 17, 2026. Claude Fable 5.2 came out Nov 9, 2026; Step 5 Preview came out Sep 16, 2026. Knowledge cutoff: GPT-6.2 Astra May 18, 2026, Claude Fable 5.2 Jun 2026.

How do you decide the winner?

This rating uses price (50%, blended price per million tokens (3 input : 1 output), log scale); inputs & features (30%, image, PDF, audio and video input, tool calling, structured output and reasoning); context window (20%, 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 18, 2026.