Qwen: Qwen2.5 VL 72B Instruct
Released Feb 1, 2025Ranked in:Vision
Try it on:
Qwen2.5-VL is proficient in recognizing common objects such as flowers, birds, fish, and insects. It is also highly capable of analyzing texts, charts, icons, graphics, and layouts within images.
Best overall rank
#95
Across all subsets · overall
Total arena votes
6,056
Across 3 (subset, category) entries
Providers on OR
1
Parasail
First seen
2026-05-08
45 days tracked
Trend
Pricing
Per token
- Input (per 1M)
- $0.800
- Output (per 1M)
- $1.00
- Cache Read (per 1M)
- $0.400
Capabilities
Modality
text+image->textContext
131K tokens
Tokenizer
Qwen
Knowledge cutoff
2024-06-30
Max output
128K tokens
Moderation
Unmoderated
Hugging Face
Supported parameters
frequency_penalty
logit_bias
logprobs
max_tokens
presence_penalty
repetition_penalty
response_format
seed
stop
structured_outputs
temperature
top_k
top_logprobs
top_p
Providers
| Provider | Region / Tag | Context | Input/M | Output/M | Uptime 1d | Latency | Quant | Cache |
|---|---|---|---|---|---|---|---|---|
| Parasail | parasail/fp8 | 128K | $0.800 | $1.00 | 100.00% | - | fp8 | - |
Per-category ranks
Vision
| Rank | Category | Variant | Trend | Elo | Votes |
|---|---|---|---|---|---|
| #74 | Chinese | - | 1121 | 119 | |
| #95 | English | - | 1125 | 2,169 | |
| #95 | Overall | - | 1122 | 3,768 |