Models head to head — pick two or three. Same scoring, same week. Search to swap, add, or remove a model and the matchup recomputes. Hover or tap a stat to see how it shifted.
If quality matters most
OGPT-5
Best raw competency at 58.5/100 and the lowest hallucination rate in the set. Pay the premium when accuracy compounds.
If cost matters most
AQwen 3 Max
91% cheaper than GPT-5, 10.5 points lower on competency. The value pick for high-volume workloads.
OpenAIOGPT-5
Cloud · API
AlibabaAQwen 3 Max
Open weights
CompetencyWeighted LLM score
58.5/100
flat W21 · rank 2
Best in class
48/100
flat W21 · rank 89 · −10.5 vs winner
Cost (input)Per 1M tokens
$14.00
~12× the cheapest in the row
$1.20
91% cheaper than GPT-5
Cheapest
Cost (output)Per 1M tokens
$42.00
~7× the cheapest in the row
$6.00
86% cheaper
Cheapest
SpeedTokens / second
76 tok/s
local inference, batch 1
Fastest cloud
0 tok/s
TTFT 0s
AccuracyHallucination rate
2.1%
lowest in category
Most accurate
—
no data
ContextMax input tokens
~256 pages
~256 pages context window
Largest
~256 pages
~256 pages context window
DeploymentWhere it runs
Cloud only
API key · quick setup
Cloud or local
Self-host on your GPUs
Most flexible
Data residencyCompliance
Vendor regions
verify region availability
Anywhere
runs inside your perimeter
Most flexible
LicenseWhat you can do
Proprietary
commercial OK · ToS applies
Open weights
commercial OK
Most permissive
Where each one wins.
Sub-competency scores by business task. Bars highlight the row leader in accent.
Customer service40% of category weight
OGPT-5
96
AQwen 3 Max
81
Research & analysis22% weight
OGPT-5
93
AQwen 3 Max
78
Summarization18% weight
OGPT-5
92
AQwen 3 Max
77
Code review10% weight
OGPT-5
89
AQwen 3 Max
74
Document understanding6% weight
OGPT-5
86
AQwen 3 Max
71
OGPT-5
✓ Best for
Strong at Customer Service for OpenAI.
Strong at Content Writing for OpenAI.
⚠ Watch out for
Validate fit for your specific workload before committing.
