Qwen 3.8 27B
qwen/qwen3.8-27b
compact vision-language coding & agentic work.
cmd --model qwen/qwen3.8-27b
Intelligence index
52
Output speed
50.5 tok/s
Input
$0.40 /M
Output
$3 /M
Cache read
$0.04 /M
Agent-loop cost
$0.15 /M in
Context window
262K tokens
Released
August 14, 2026
Modalities
→
vs. the lineup
Qwen 3.8 27B beside its stablemates and nearest rivals. The ◆ marks the best value in each column across every row shown.
pin a rival:
| Model | Intelligence | Coding | Speed | Input $/M | Output $/M | Blended $/M | Context |
|---|---|---|---|---|---|---|---|
| Qwen 3.8 Max | 58.1◆ | 71.8◆ | 23.6◆ | $2◆ | $6◆ | $3◆ | 1M◆ |
| Qwen 3.8 27B ◆ | 52◆ | 68.1◆ | 50.5◆ | $0.40◆ | $3◆ | $1.05◆ | 262K◆ |
| Qwen 3.7 Max | 46.7◆ | 66◆ | —◆ | $2.50◆ | $7.50◆ | $3.75◆ | 1M◆ |
| Qwen 3.7 Plus | 39.4◆ | 55.9◆ | 55.5◆ | $0.40◆ | $1.60◆ | $0.70◆ | 1M◆ |
| Qwen 3.8 Max 0902 | —◆ | —◆ | —◆ | $2◆ | $6◆ | $3◆ | 1M◆ |
| Qwen 3.8 Flash | —◆ | —◆ | —◆ | $0.16◆ | $0.47◆ | $0.24◆ | 1M◆ |
| DeepSeek V4 Pro (latest)pinned | 53.2◆ | 68.8◆ | 61◆ | $0.66◆ | $1.98◆ | $0.99◆ | 1M◆ |
| DeepSeek V4 Flash (latest)pinned | 52◆ | 69.1◆ | 129◆ | $0.22◆ | $0.66◆ | $0.33◆ | 1M◆ |
coding performance
The Intelligence Index and its sub-scores, ranked against every scored model in the catalog. A metric that has not been measured for Qwen 3.8 27B has been left empty.
Coding Index
68.1
#26 of 46 scored
Terminal-Bench
79.8
#19 of 46 scored
Intelligence Index
52
#23 of 52 scored
Long-context reasoning
77.3
reasoning across a long context
SciCode
44.7
scientific coding
GPQA Diamond
90.5
graduate-level QA
usage calculator
How far a month of credits goes on Qwen 3.8 27B.
Input tokensfresh prompt
800Output tokensmodel reply
180Cache read tokensre-read context
50Kcost / request $0.0029 · in $0.40 · out $3 · cache $0.04 per M
fresh input 11%output 19%cache reads 70%
Requests / 30 days
3.5K
$10 credits ÷ $0.0029 per request
~699 quick fixes~140 bug fixes~23 feature PRs
what real work costs
Real coding tasks priced end to end on Qwen 3.8 27B, from a quick lookup to a full-repo agent run.
One agent task
$0.06
180K in at 75% cache hit, 12K out
What you pay
$0.31 /M
all-in across every token that task touched
Sticker input
$0.40 /M
cache reads bill at $0.04 /M instead
| Task | Tokens in · out | Qwen 3.8 27B | GLM-5.3 Flash | Claude Haiku 4.5 |
|---|---|---|---|---|
| Quick lookup / one-liner | 8K · 1K | $0.0036 | $0.0008 | $0.0065 |
| Review a 500-line PR | 60K · 4K | $0.02 | $0.0060 | $0.04 |
| Fix a bug (agent loop) | 180K · 12K | $0.06 | $0.02 | $0.12 |
| Refactor a module | 320K · 20K | $0.11 | $0.03 | $0.21 |
| Full-repo agent run | 900K · 45K | $0.24 | $0.07 | $0.49 |
frequently asked
How much does Qwen 3.8 27B cost?
$0.40/M input and $3/M output, cache reads $0.04/M. In an agent loop most input is cache-read, so the effective input rate is about $0.15/M.Which plan do I need?
Available on Go and above.
How do I switch to it?
Run
cmd --model qwen/qwen3.8-27b, or type /model in a session and pick it. You can switch mid-session without losing context.Ship code that matches your taste
Command Code is the AI coding agent that continuously learns your taste. Start for $1.
Benchmarks from Artificial Analysis (v4.1)commandcode.ai/models/qwen3-8-27b