Shanghai AI Lab

InternVLA-A1.5↗

InternVLA-A1.5 is a VLA from Shanghai AI Lab, released July 6, 2026. It has no overall rank on the Robot Intelligence Index, which needs both a simulation and a real-robot result. Its best result is #10 of 15 on EBench. Shanghai AI Lab VLA unifying understanding, latent foresight tokens, and flow-matching actions on a Qwen3.5-2B backbone.

Release date: Jul 06, 2026

Type
VLA
Model weights
Open
License
Non-commercial
Parameters
—
Pricing
Self-hosted
Benchmarks
2
Overall rank
Not ranked
Needs a real-robot result to rank overall.
Real robots rank
Not ranked
No real-robot results yet.
Simulation rank
#27of 49
Rating 1364 · range 19–36
Rating
1397± 20.4
10601989
Cost / task · est.
$0.032
$0.0087$17.74
Time / task · est.
35s
9.5 s1 h 22 min
Rank on each benchmark
BenchmarkScoreRank
EBenchSuccess rate34.210 / 15RoboColiseumTotal0.4726 / 30

Per task attempt; the tiles show the median across benchmarks. Self-hosted compute is priced at one H100 at $3.29 per hour (Lambda on-demand, Oct 2026).

Other models from Shanghai AI Lab

Inference

No latency published; Qwen3.5-2B VLM backbone plus lightweight action expert (~460M per secondary summary); video branch dropped at inference.