The Model Index
Up to four models

GPT-6.2 Astra vs Claude Opus 5.5

On the index (GPT-4 = 100), GPT-6.2 Astra scores 202 and Claude Opus 5.5 195. Side by side on the same third-party tests, with price and context. The link in your address bar shares this exact comparison.

Next to GPT-6.2 Astra
Presets
GPT-6.2 Astra
202184–220
WeightsClosed
ReleasedNov 2026
Price in / out$8.0 / $40
Context1.05M
Core tests taken3 of 9
Claude Opus 5.5
195178–211
WeightsClosed
ReleasedSep 2026
Price in / out$4.0 / $20
Context1M
Core tests taken6 of 9

GPT-6.2 Astra and Claude Opus 5.5 are too close to call (7 points apart, within the uncertainty), ahead on 2 of 3 shared core tests, at 2.0× the price.

Test by testthird-party results

GPT-6.2 AstraClaude Opus 5.5
0%25%50%75%100%REASONING & KNOWLEDGEGPQA Diamond (Epoch AI)GPQA Diamond (Epoch AI)EBR-benchEBR-benchSimpleQA VerifiedSimpleQA VerifiedMATHSFrontierMath (tiers 1-3)FrontierMath (tiers 1-3)AIME-style maths (OTIS mock)AIME-style maths (OTIS mock)Chess puzzlesChess puzzlesFrontierMath (tier 4)FrontierMath (tier 4)CODINGTerminal-Bench 4.0Terminal-Bench 4.0Vibe Code Bench v1.1Vibe Code Bench v1.1MirrorCodeMirrorCodeAGENTSBioMysteryBenchBioMysteryBenchTerminal-Bench-ScienceTerminal-Bench-ScienceFurniture assemblyFurniture assemblyMystery game puzzlesMystery game puzzlesPROFESSIONAL WORKHLAB-Vals-1.1HLAB-Vals-1.1Legal Research BenchLegal Research BenchTax Agent BenchTax Agent BenchNOVEL PROBLEMSMysteryMechanismMysteryMechanism
Every result as a table
TestGPT-6.2 AstraClaude Opus 5.5
Reasoning & knowledge
GPQA Diamond (Epoch AI)96.8%94.1%
EBR-bench92.4%71.4%
SimpleQA Verified78.5%72.2%
Maths
FrontierMath (tiers 1-3)92.6%91.2%
AIME-style maths (OTIS mock)100.0%100.0%
Chess puzzles78.0%67.0%
FrontierMath (tier 4)100.0%95.0%
Coding
Terminal-Bench 4.063.1%65.1%
Vibe Code Bench v1.189.9%90.3%
MirrorCode67.8%95.6%
Agents
BioMysteryBench80.7%79.3%
Terminal-Bench-Science71.4%47.1%
Furniture assembly89.2%83.3%
Mystery game puzzles95.0%71.0%
Professional work
HLAB-Vals-1.116.3%9.2%
Legal Research Bench42.8%50.5%
Tax Agent Bench21.8%44.5%
Novel problems
MysteryMechanism53.1%49.5%

Tests at least two of these models have taken. Bold rows are the core tests.

This page as a chartDownload PNG