SpaceXSI
Grok 4.3
Closed weights
- Intelligence
- 14.0
- Rank 326 of 744 measured models
- Coding
- 35.2
- Rank 171 of 298 measured models
- Output speed
- 93.9 t/s
- Rank 184 of 403 measured models
- Price/1M
- $1.56
- Rank 441 of 659 measured models
- Context window
- 1M
- Rank 84 of 1199 measured models
- Cost per task
- $0.17
- Rank 67 of 193 measured models
Grok 4.3 (Non-reasoning) scores 14.0 on the Artificial Analysis Intelligence Index, ranking 326th of 744 measured models. The closest models on this index are Gemma 4 12B (14.2), Grok Code Fast 1 (14.1), and DeepSeek V3.1 Terminus (13.9). At $1.56 per million tokens (3:1 input/output blended price), it is more expensive than Gemma 4 12B ($0.15) and DeepSeek V3.1 Terminus ($0.453). With an output speed of 93.9 t/s, it is faster than DeepSeek V3.1 Terminus (40.8 t/s), slower than Gemma 4 12B (112 t/s).
Models at a similar level
The closest models on the Intelligence Index, one configuration per family.
- Gemma 4 12B
Google · $0.15 · 112 t/s
14.2+0.2$0.15112 t/sCompare - Mistral Medium 3.5
Mistral · $3 · 151 t/s
14.2+0.2$3151 t/sCompare - Grok Code Fast 1
SpaceXSI · — · —
14.1+0.1——Compare - DeepSeek V3.1 Terminus
DeepSeek · $0.453 · 40.8 t/s
13.9−0.1$0.45340.8 t/sCompare
