InternVL3 5 4B Instruct measured on 2 of 15 Vision & Documents benchmarks2
BenchmarkByScoreRank
Legacy · saturated or no longer reported2
No benchmark matches
—per task?
self-host onlyprice
—output speed?
41Kcontext
Alternatives?
ModelScoreCostPer taskWhat changes
BetterGemini 3 Flash Previewunstated×3.3—$0.50 / $3
OpenQwen3 VL 8B Instructunstated+25%—$0.18 / $0.69Memory (4-bit): 2.8 GB → 4.4 GB
BetterMiMo v2 Omniunstated×3—$0.14 / $0.28
BetterKimi K2.5unstatedOpen×2.8—$0.60 / $3Memory (4-bit): 2.8 GB → 616 GB
BetterGPT-5unstated×2.7—$1.25 / $10ModelScoreCostPer taskWhat changes
Nothing here yet.
ModelScoreCostPer taskWhat changes
Gemini 3 Flash Previewunstated×3.3—$0.50 / $3MiMo v2 Omniunstated×3—$0.14 / $0.28Kimi K2.5unstatedOpen×2.8—$0.60 / $3Memory (4-bit): 2.8 GB → 616 GBGPT-5unstated×2.7—$1.25 / $10Gemma 4 31BunstatedOpen+92%—$0.10 / $0.34Memory (4-bit): 2.8 GB → 19 GBModelScoreCostPer taskWhat changes
Qwen3 VL 8B InstructunstatedOpen+25%—$0.18 / $0.69Memory (4-bit): 2.8 GB → 4.4 GBMiMo VL 7B RL 2508Open+20%——Memory (4-bit): 2.8 GB → 4.9 GBQwen3 VL 4B InstructunstatedOpen+17%—$0.10 / $0.60OInternVL3 5 8B InstructOpen+15%——Memory (4-bit): 2.8 GB → 5 GBMiMo VL 7B SFT 2508Opensame——Memory (4-bit): 2.8 GB → 4.9 GBModelScoreCostPer taskWhat changes
OInternVL3 5 241B A28B InstructOpen+60%——Memory (4-bit): 2.8 GB → 141 GBOInternVL3 5 30B A3B InstructOpen+27%——Memory (4-bit): 2.8 GB → 18 GBOInternVL3 5 8B InstructOpen+15%——Memory (4-bit): 2.8 GB → 5 GBOInternVL3 1BOpenn/a——Context: 41K → 33KOInternVL3 2BOpenn/a——Context: 41K → 33KOInternVL3 8BOpenn/a——Context: 41K → 33K+1All OpenGVLab models →Run it yourself
estimateFits on✓ 8 GB (phone, laptop) · 8-bit✓ 16 GB · full quality✓ 24 GB GPU · full quality✓ 48 GB GPU · full quality✓ 80 GB GPU · full quality
LicenceApache-2.0free for commercial use
Benchmarks
in Vision & DocumentsVideo-MME 2 · with subtitles/audio14.2#17
Video-MME 2 · without subtitles/audio9.9#16
FAQ
How much does InternVL3 5 4B Instruct cost?›
There is no per-unit API price: self-host only.
Is InternVL3 5 4B Instruct free to use?›
There is no free API. The weights are open, so you can run it yourself.
How good is InternVL3 5 4B Instruct, and what is it best at?›
#14 of 16 in Vision & Documents.
Is InternVL3 5 4B Instruct better than MiMo VL 7B SFT 2508?›
On Video-MME 2 · without subtitles/audio, InternVL3 5 4B Instruct scores 9.9 vs 10.7 — about the same.
What is the context window of InternVL3 5 4B Instruct?›
41K tokens.
When was InternVL3 5 4B Instruct released? Is there a newer version?›
Released 2025-08-25. It is the current version in its line.
Is InternVL3 5 4B Instruct open source?›
Yes, the weights are open (Apache-2.0 licence).
Can I run InternVL3 5 4B Instruct locally?›
The weights are open and can be self-hosted; hardware needs are not confirmed yet.
Can InternVL3 5 4B Instruct read images?›
It accepts text and images as input.
What is InternVL3 5 4B Instruct?›
InternVL3 5 4B Instruct is an open-weights language model from OpenGVLab, released Aug 2025.
What are the best alternatives to InternVL3 5 4B Instruct?›
Gemini 3 Flash Preview (better), Qwen3 VL 8B Instruct (open weights).