Compare · ModelsLive · 2 picked · head to head

Claude Sonnet 5.5 vs GPT-6.2 Astra

Side by side · benchmarks, pricing, and signals you can act on.

Winner summary

GPT-6.2 Astra wins 8 of 10 shared benchmarks. Leads in knowledge · math · general.

Category leads
knowledge·GPT-6.2 Astramath·GPT-6.2 Astrageneral·GPT-6.2 Astracoding·Claude Sonnet 5.5

Head to head · from the data · data as of 2026-11-18

  • Benchmarks · Claude Sonnet 5.5 leads on 1 of 10 shared benchmarks (Mirrorcode), GPT-6.2 Astra on 8 (Chess Puzzles, FrontierMath-Tier-4-v2-Private, FrontierMath-Tiers-1-3-v2-Private, Furniture Assembly, 4 more) · 1 tied
  • Price · Claude Sonnet 5.5 input $2.00/M vs GPT-6.2 Astra $8.00/M, Claude Sonnet 5.5 is 75% cheaper (OpenRouter list price, as of 2026-11-18)
  • Cheapest provider · Claude Sonnet 5.5 at Amazon Bedrock and 4 other providers $2.00/M input (of 10 offers) as of 2026-11-18 · GPT-6.2 Astra at Amazon Bedrock and 2 other providers $8.00/M input (of 8 offers) as of 2026-11-18
  • Context · Claude Sonnet 5.5 1.0M tokens vs GPT-6.2 Astra 1.1M tokens
  • Released · Claude Sonnet 5.5 Sep 28, 2026 vs GPT-6.2 Astra Nov 17, 2026 · 2 months apart
  • Price history · Claude Sonnet 5.5 OpenRouter input list price had no move of 5% or more in the 90 days to 2026-11-18 · GPT-6.2 Astra OpenRouter input list price had no move of 5% or more in the 90 days to 2026-11-18

Sources · scores from public leaderboards, list prices and provider offers from OpenRouter, price moves from the BenchGecko changelog (5% or more, held in the next daily check), Gecko tests run by BenchGecko.

Hype vs Reality
Claude Sonnet 5.5
#19 by perf·#12 by attention
UNDERRATED
GPT-6.2 Astra
#1 by perf·#2 by attention
DESERVED
Best value
3.4x better value than GPT-6.2 Astra
Claude Sonnet 5.5
11.9 pts/$
$6.00/M
GPT-6.2 Astra
3.5 pts/$
$24.00/M
Vendor risk
Anthropic logo
Anthropic
$1.40T·Tier 1
Medium risk
OpenAI logo
OpenAI
$840.0B·Tier 1
Medium risk
Head to head
Claude Sonnet 5.5GPT-6.2 Astra
Chess Puzzles
GPT-6.2 Astra leads by +20.0
Chess Puzzles · tests strategic and tactical reasoning by having models solve chess puzzle positions, evaluating lookahead and pattern recognition abilities.
Claude Sonnet 5.5
56.9
GPT-6.2 Astra
76.8
FrontierMath-Tier-4-v2-Private
GPT-6.2 Astra leads by +19.5
Claude Sonnet 5.5
80.5
GPT-6.2 Astra
100.0
FrontierMath-Tiers-1-3-v2-Private
GPT-6.2 Astra leads by +4.1
Claude Sonnet 5.5
93.4
GPT-6.2 Astra
97.5
Furniture Assembly
GPT-6.2 Astra leads by +20.2
Claude Sonnet 5.5
64.3
GPT-6.2 Astra
84.5
GPQA diamond
GPT-6.2 Astra leads by +1.8
Graduate-Level Google-Proof QA (Diamond set) · expert-crafted questions in physics, biology, and chemistry that are difficult even for domain PhDs.
Claude Sonnet 5.5
98.0
GPT-6.2 Astra
99.8
Mirrorcode
Claude Sonnet 5.5 leads by +30.0
Claude Sonnet 5.5
97.8
GPT-6.2 Astra
67.8
Mystery Game Puzzles
GPT-6.2 Astra leads by +33.0
Claude Sonnet 5.5
61.5
GPT-6.2 Astra
94.5
OTIS Mock AIME 2024-2025
OTIS Mock AIME 2024-2025 · simulated American Invitational Mathematics Examination problems testing advanced problem-solving skills.
Claude Sonnet 5.5
100.0
GPT-6.2 Astra
100.0
SimpleQA Verified
GPT-6.2 Astra leads by +32.0
SimpleQA Verified · short factual questions with verified answers, measuring factual accuracy and the tendency to hallucinate or provide incorrect information.
Claude Sonnet 5.5
46.5
GPT-6.2 Astra
78.5
SpireClimb
GPT-6.2 Astra leads by +40.0
Claude Sonnet 5.5
1.3
GPT-6.2 Astra
41.3
Full benchmark table
BenchmarkClaude Sonnet 5.5GPT-6.2 Astra
Chess Puzzles
Chess Puzzles · tests strategic and tactical reasoning by having models solve chess puzzle positions, evaluating lookahead and pattern recognition abilities.
56.976.8
FrontierMath-Tier-4-v2-Private
80.5100.0
FrontierMath-Tiers-1-3-v2-Private
93.497.5
Furniture Assembly
64.384.5
GPQA diamond
Graduate-Level Google-Proof QA (Diamond set) · expert-crafted questions in physics, biology, and chemistry that are difficult even for domain PhDs.
98.099.8
Mirrorcode
97.867.8
Mystery Game Puzzles
61.594.5
OTIS Mock AIME 2024-2025
OTIS Mock AIME 2024-2025 · simulated American Invitational Mathematics Examination problems testing advanced problem-solving skills.
100.0100.0
SimpleQA Verified
SimpleQA Verified · short factual questions with verified answers, measuring factual accuracy and the tendency to hallucinate or provide incorrect information.
46.578.5
SpireClimb
1.341.3
Pricing · per 1M tokens · projected $/mo at 10M tokens
ModelInputOutputContextProjected $/mo
Anthropic logoClaude Sonnet 5.5$2.00$10.001.0M tokens (~500 books)$40.00
OpenAI logoGPT-6.2 Astra$8.00$40.001.1M tokens (~525 books)$160.00
Claude Sonnet 5.5 leads on 1 of 10 shared benchmarks (Mirrorcode), GPT-6.2 Astra on 8 (Chess Puzzles, FrontierMath-Tier-4-v2-Private, FrontierMath-Tiers-1-3-v2-Private, Furniture Assembly, 4 more) · 1 tied