Qwen 2.5 7B Instruct Turbo
An instruction-tuned causal language model of 7.61B parameters in the Qwen2.5 series, with improved coding, mathematics, long-text generation, structured data understanding and JSON output across 29+ languages.
text → text · made by Alibaba
Sold by 1 way
| Seller | Lane | Rate |
|---|---|---|
| Together AI aggregator | standard | $0.3per Mtok in$0.3per Mtok outdocs.together.ai · read 2026-08-24 |
Measured 1 standing
| Place | Board | Metric | Score |
|---|---|---|---|
| 4thof 25 | AgentBench Leaderboard (new) | Success Rate (pass@1) AVG | 62 |
About
Qwen 2.5 7B Instruct Turbo — a text model from Alibaba, sold by one company from $0.3 in and $0.3 out per million tokens, placed 4th of 25 on AgentBench Leaderboard (new).
It takes text and returns text, with a context window of 32,768 tokens. It was published in September 2024. Its sellers say it can call a tool. The catalogue files it under agents. Only Together AI sells it, at $0.3 in and $0.3 out per million tokens. It stands 4th of 25 on AgentBench Leaderboard (new).
Every current figure
- Maker
- Alibaba
- Register
- model
- Takes
- text
- Returns
- text
- Context
- 32,768 tokens
- Published
- September 2024
- Parameters
- 7 billion · read from its own name
- Sellers
- 1
- Maker's own price
- not read
- Price
- $0.3 in and $0.3 out per million tokens — Together AI
- Boards
- 1
- Best place
- 4th of 25 — AgentBench Leaderboard (new)
Known as 2 names
AgentRL w/ Qwen2.5-7B-InstructQwen/Qwen2.5-7B-Instruct-Turbo