By the toolist.ai editors. Researched from public sources, not hands-on tested by toolist.ai. Updated November 19, 2026; prices checked November 19, 2026.

GPT-6.2 Astra vs Claude Fable 5.2

Verdict

Both cost $8 per million input tokens and $40 per million output tokens, so pick on the work. Claude Fable 5.2 leads on the Artificial Analysis Intelligence Index, 65 against 62 at max effort and 62 against 57 at everyday effort (high for Fable 5.2, medium for Astra), and on HLE-Diamond (80% against 72%). GPT-6.2 Astra leads both terminal benchmarks: 64% against 60% on Terminal-Bench 4.0 and 78% against 73% on Terminal-Bench-Science. Fable 5.2 also measures more than twice the output speed, 107.4 against 46.8 tokens a second, and its cached input costs $0.20 against $0.40.

Verdict based on published prices, limits and benchmarks, not on our own tests.

Choose GPT-6.2 Astra if

  • Terminal and science agents: 78% against 73% on Terminal-Bench-Science and 64% against 60% on Terminal-Bench 4.0
  • You already build on the OpenAI API, or use ChatGPT Work and Codex on a Plus, Pro, Business, Enterprise or Edu plan
  • You want a slightly larger context window: 1,050,000 tokens against 1,000,000

Choose Claude Fable 5.2 if

  • Hard reasoning and research: 80% against 72% on HLE-Diamond
  • You want faster output: 107.4 against 46.8 tokens a second at each model's everyday effort
  • Prompt-heavy agents: cached input costs $0.20 per million tokens against $0.40
  • You build on Amazon Bedrock or Google Cloud, where Anthropic offers Fable 5.2

Price

Input costs the same on both. Output costs the same on both.

Per million tokensGPT-6.2 AstraClaude Fable 5.2
Input$8 source$8 source
Cached input$0.40 source$0.20 source
Output$40 source$40 source

GPT-6.2 Astra: Cache writes cost $10 per million tokens with 30-minute retention and $20 with 12-hour retention. Prompts over 272,000 input tokens cost twice the input and cache rates and 1.5 times the output rate for the whole request. Batch costs half and Fast mode double; unlike GPT-6 Astra, there is no Flex tier. Ultrafast costs three times the standard rate through February 28, 2027 and six times from March 1, 2027. Eligible background requests cost 30% less with a two-hour completion window and 40% less with an eight-hour one.

Claude Fable 5.2: Cache writes cost $10 per million tokens (5-minute cache) or $16 (1-hour cache). The Batch API costs half. There is no long-context surcharge. Fast mode, with up to 2.5 times faster output, costs $16 in and $80 out and is not available with the Batch API.

Worked example: what a month costs

We priced three monthly workloads at the standard rates. The provider does not publish these figures.

WorkloadGPT-6.2 AstraClaude Fable 5.2
Support chatbot
100,000 replies, each 3,000 input and 400 output tokens, no caching (300M in, 40M out)
$4,000$4,000
Document Q&A
5,000 questions over the same 60,000-token document set (90% cache hits), 800 output tokens each (300M in, 4M out)
$508$454
Coding agent
200 tasks, each 1.5 million input tokens (90% cache hits) and 50,000 output tokens (300M in, 10M out)
$748$694

Price your own workload

ModelCost a month
GPT-6.2 Astra $600
Claude Fable 5.2 $600

Standard API prices, no batch or priority discounts. Cached input is billed at the cache-read price; cache-write charges are not included. Reasoning models bill their thinking as output tokens.

Limits and release dates

GPT-6.2 AstraClaude Fable 5.2
ReleasedNovember 17, 2026 sourceNovember 9, 2026 source
Context window1,050,000 tokens source1,000,000 tokens source
Maximum output128,000 tokens source128,000 tokens source
Knowledge cutoffMay 18, 2026 sourceJune 2026 source
Input and outputText and images in; text out sourceText and images in; text out source
API model namegpt-6.2-astraclaude-fable-5-2

Benchmarks

Scores as published by the provider or measured by Artificial Analysis; higher is better. A dash means no score was published.

BenchmarkGPT-6.2 AstraClaude Fable 5.2
Artificial Analysis Intelligence Index
max effort
62.0 source65.0 source
Intelligence Index at everyday effort 57.0 (medium effort, where OpenAI suggests starting) source62.0 (high effort, the API default) source
Terminal-Bench 4.0
max effort
64.0% source60.0% source
HLE-Diamond
max effort
72.0% source80.0% source
Terminal-Bench-Science 0.1
max effort
78.0% source73.0% source
Output speed 46.8 tokens/s (medium effort, Artificial Analysis) source107.4 tokens/s (high effort, Artificial Analysis) source

Benchmark and speed data from Artificial Analysis where linked.

Read more

Sources