SShogo Bench

Claude Sonnet 5.5 vs GPT-6.1 Terra

Claude Sonnet 5.5 leads GPT-6.1 Terra by 0.0 points on the Shogo Agent Index, but their ranges overlap, so treat it as close. Claude Sonnet 5.5 wins 7 and GPT-6.1 Terra wins 4 of 12 shared benchmarks. GPT-6.1 Terra is cheaper for agent workloads.

Add more models to this comparison →

Pillars

PillarClaude Sonnet 5.5GPT-6.1 Terra
Agentic coding7979
Long-horizon autonomy7574
Knowledge-work agents5756
Computer use6163
Security & context learning6162

Shared benchmarks

BenchmarkClaude Sonnet 5.5GPT-6.1 Terra
APEX-Agentsindex75.5%68.3%
Vending-Bench 2index$11,165$9,107
CritPt31.4%28.0%
FrontierMath Tier 480.5%92.7%
FrontierMath Tiers 1-388.8%90.9%
GPQA Diamond95.6%94.3%
MirrorCode97.8%76.7%
OTIS Mock AIME100.0%100.0%
SciCode61.0%55.1%
SimpleBench75.9%72.2%
SimpleQA Verified46.5%57.3%
WeirdML v319.9%28.1%

Specs and price

Claude Sonnet 5.5GPT-6.1 Terra
Agent Index69 (#9)69 (#8)
Agent $/M$0.84$0.42
Input / output $/M$2.00 / $10.00$1.00 / $5.00
Context1M1.1M
ReleasedSep 28, 2026Oct 26, 2026
WeightsClosedClosed
API hosts54

Don’t pick one. Route.

Shogo can run Claude Sonnet 5.5 and GPT-6.1 Terra side by side and send each step of an agent to the one that’s cheaper for it.

Try both in Shogo