Skip to content

All Models

Rank #13 / 13

Molmo 2 8B

Molmo 2 8B is Ai2's Molmo model (open weights). It is ranked 13 of 13 models with enough coverage on the Visual Intelligence Index.

  • Ai2
  • Open weights
  • Molmo
  • Video input
  • Structured output
  • Tool calling
  • Streaming
  • Self-hostable
Read the report
53.3%Visual Intelligence Index · Rank #13 of 13

Capability profile

Comparison Summary

Molmo 2 8B ranks 13 of 13 on the perception index. Immediately ahead is Nemotron 3 Nano Omni 30B-A3B. List price is $0.10 per million input tokens and $0.40 per million output. 0 of 13 listed models cost less to prompt. Latency is 0.2s with a 128K context window and tokens/s. Prices and most scores in this snapshot are placeholders, not sourced figures. Strongest capability in this set is Causal Reasoning (62%). Weakest is Action / Event Understanding (50%).

Visual Intelligence
57.5%
Availability
Open weights
Input / 1M tokens
$0.10
Output / 1M tokens
$0.40
Latency
0.2s
Context Window
128K
Speed (tokens/s)
Parameters
8B
Time to first token

All benchmark scores

BenchmarkMetricScore95% CISettingSource
Video-MME v2Accuracy, no subtitles27.5%26.0–29.11 fps, 338-frame cap (673 videos thinned)Reported
LVBenchTest accuracy48.6%46.1–51.11 fps, 338-frame cap (103 videos thinned)Reported
Perception TestOverall accuracy67.4%63.1–71.71 fpsReported
NExT-QAAccuracy (hard split)89.2%88.4–90.11 fpsReported
Q-Bench VideoOverall accuracy62.9%59.7–66.11 fpsReported
EgoSchemaAccuracy (fullset)58.2%53.9–62.51 fpsReported
UCF101-ADAccuracy19.5%18.3–20.71 fpsReported

Benchmark Scores

Compare reported model scores across each available benchmark or capability index.

13 of 13 models
Filter by model access

Display

Capabilities Index Scores

Compare reported model scores across each available benchmark or capability index.

13 of 13 models
Filter by model access

Display

Usability

3.5/ 5

What it's like to build against this model, scored out of five from the developer-facing capabilities in the dataset.

Reachable0.5

There is an endpoint you can call without hosting anything.

  • First-party API — supported
  • Third-party API — not supported
Portable1.0

You can run it yourself, and are not tied to one vendor.

  • Open weights — supported
  • Self-hostable — supported
Multimodal input0.5

It takes the footage directly, rather than frames you extracted.

  • Video ingestion — supported
  • Audio — not supported
Programmable1.0

Output you can parse, and tools it can call on its own.

  • Structured output — supported
  • Tool calling — supported
Operable0.5

Usable interactively and in bulk, not only one call at a time.

  • Streaming — supported
  • Batch API — not supported

Other Ai2 models

Other models from the Ai2 family.

DateModelVisual IntelligenceLatency
Apr 2, 2026Molmo 2 8B57.5%0.2s