Qwen2.5-VL-72B-Instruct: benchmark scores, price and ranking

Qwen2.5-VL-72B-Instruct from Qwen ranks #418 of 743 LLMs on the LLMs Tiger aggregate benchmark score (47.6), based on 2 benchmark results captured up to 2026-10-13.

Organisation: Qwen
Type: Open-weight / local model
Open weights: Yes
Released: 2024-09-19 (source: epoch)
Cheapest API price: $0.250 in / $0.750 out per million tokens (via nebius, observed 2026-10-13T20:47Z)
Aggregate rank: #418 of 743 (score 47.6)

Across the 1 benchmark where at least ten models are scored, Qwen2.5-VL-72B-Instruct is in the top 10% on 0 and in the top half on 1.

Its cheapest listed input price, $0.250 per million tokens, is lower than that of 64% of the 427 models with a listed price.

Qwen2.5-VL-72B-Instruct benchmark results

Qwen2.5-VL-72B-Instruct benchmark scores, ranked against all models measured on each benchmark
Benchmark	Score	Rank	Source	Captured
MMLU-Pro	64.9%	#52 of 136	HuggingFace MMLU-Pro leaderboard	2026-10-13
OSWorld	5.0%	#5 of 5	Epoch AI Benchmarking Hub	2026-10-13

About Qwen2.5-VL-72B-Instruct

Qwen2.5-VL-72B-Instruct is an open-weight model that can run on your own hardware from Qwen. It was released on 2024-09-19, according to Epoch AI. The cheapest API listing we track charges $0.250 per million input tokens and $0.750 per million output tokens, through nebius. We hold 2 benchmark results for it across 2 areas, and none is in the top tenth of its board.

Where Qwen2.5-VL-72B-Instruct stands

On our aggregate score, Qwen2.5-VL-72B-Instruct is #418 of 743 ranked models, around the middle of the table. Its score is 47.6; the scale is centred on 50, so it sits 2.4 points below the typical model. That beats 44% of the ranked field. Among the 23 ranked Qwen models it is #13. The next one up is Qwen2.5 (7B + 72B) at 48.2. The next one down is Qwen1.5-110B at 47.1.

Qwen2.5-VL-72B-Instruct, benchmark by benchmark

MMLU-Pro puts Qwen2.5-VL-72B-Instruct at 64.9%, #52 of 136. That beats 62% of the models tested, but 23.2 points short of Darwin-180B-RSI (88.1%). The results either side are Seed-OSS-36B-Base (65.1%) and DeepSeek-V3 (64.4%). It covers 57-subject knowledge + reasoning. The board median is 55.3%, so Qwen2.5-VL-72B-Instruct is 9.7 points above it. Source: HuggingFace MMLU-Pro leaderboard, captured 2026-10-13.

The full table above also lists the other 1 result, including boards with fewer than ten models, where a rank says little.

Strengths and weaknesses

Qwen2.5-VL-72B-Instruct has too few results on boards with ten or more models to say where it is strong or weak. Treat any single score as a data point, not a verdict.

Price and value

The median input price across the 427 priced models is $0.500 per million tokens, so Qwen2.5-VL-72B-Instruct costs less than the typical listing. Models that score within 1.5 points or better and cost less: Qwen2.5-Coder-7B-Instruct ($0.010, 49.5), DeepSeek V4.1 Flash (high) ($0.010, 48.9), Qwen2.5-Coder-3B-Instruct ($0.010, 46.5). Similar scores cost more on DeepSeek-V3.2 (Thinking; Fireworks) ($0.259) and DeepSeek-V3.2 ($0.259).

Release history

The next one was Qwen2.5 Coder 32B Instruct on 2024-11-12, at 48.3.

Reading these numbers

The 2 scores on this page come from 2 sources: HuggingFace MMLU-Pro leaderboard (1), Epoch AI Benchmarking Hub (1). No official leaderboard supplied any of them, so treat them as third-party measurements. With 2 results the aggregate rank can move a lot when one new score arrives.

Models ranked nearby

#415 Gemini 2.5 Flash (Preview) (Thinking 24K) (47.6)
#416 Qwen2-VL-72B-Instruct (47.6)
#417 GPT-5.5 Instant (47.6)
#419 GPT-5.1 (Thinking, Low) (47.5)
#420 grok-4-1 (47.5)
#421 Claude Code + Sonnet 5 (47.5)