TASKCLANStart building ↗

Benchmarks / T1 MAX

T1 Max.

Maximum reasoning depth for research, code and consequential work.

Routing goal: Highest quality and reasoning depth. Measured July 20, 2026, N=100 per suite.

Intelligence Index93.3#1 of 8 models measured
Blended price$10.00per 1M tokens (3:1 input:output)
Strongest suiteMMLU94% · ranked #1 of 8

Scorecard

T1 Max by benchmark.

GSM8K#5
95%Best: Kimi K2 98%
MMLUBest
94%Leads the field
MMLU-ProBest
89%Leads the field
ARC-Challenge#2
95%Best: T1 Flow 95%

Characteristics

How T1 Max behaves.

Depth & quality
Optimised for
Deliberate
Response speed
Maximum
Reasoning depth
Largest
Context handling
Highest
Relative cost
Multi-agent
Tool use
Full multimodal
Modality
Default profile

Reach for it when

Best-fit tasks.

Research & analysisCode generation & reviewStrategy & planningMultimodal creation

Compare

The other profiles.

Build with T1 Max

Ship on the right
intelligence.

Pick a T1 profile in one line and let routing choose the model. See the full comparison, or start building.

Full comparison ↗Quickstart ↗