Tencent Hy3
tencent/hy3-paid
sparse-MoE reasoning & agentic tool use.
cmd --model tencent/hy3-paid
Intelligence index
42.2
Output speed
74.8 tok/s
Input
$0.14 /M
Output
$0.58 /M
Cache read
$0.04 /M
Agent-loop cost
$0.07 /M in
Context window
262K tokens
Released
July 6, 2026
Modalities
→
vs. the lineup
Tencent Hy3 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 |
|---|---|---|---|---|---|---|---|
| Muse Spark 1.2 Contributor | 56.8◆ | 72.2◆ | —◆ | $0.10◆ | $0.20◆ | $0.13◆ | 1.05M◆ |
| MiMo V2.5 Pro | 42.9◆ | 60.2◆ | 58.1◆ | $0.43◆ | $0.87◆ | $0.54◆ | 1M◆ |
| Inkling | 42.3◆ | 52.1◆ | 69.8◆ | $1◆ | $4.05◆ | $1.76◆ | 256K◆ |
| Tencent Hy3 ◆ | 42.2◆ | 58.8◆ | 74.8◆ | $0.14◆ | $0.58◆ | $0.25◆ | 262K◆ |
| Inkling Small | 41.2◆ | 52.9◆ | 61.3◆ | $0.50◆ | $1.20◆ | $0.68◆ | 1M◆ |
| Gemini 3.5 Flash Lite | 37.4◆ | 49.3◆ | 361.2◆ | $0.30◆ | $2.50◆ | $0.85◆ | 1M◆ |
| DeepSeek V4 Pro (latest)pinned | 53.2◆ | 68.8◆ | 75.3◆ | $0.66◆ | $1.98◆ | $0.99◆ | 1M◆ |
| DeepSeek V4 Flash (latest)pinned | 52◆ | 69.1◆ | 114.6◆ | $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 Tencent Hy3 has been left empty.
Coding Index
58.8
#31 of 45 scored
Terminal-Bench
64.4
#32 of 45 scored
Intelligence Index
42.2
#34 of 51 scored
Long-context reasoning
74.7
reasoning across a long context
SciCode
47.6
scientific coding
GPQA Diamond
89.7
graduate-level QA
usage calculator
How far a month of credits goes on Tencent Hy3.
Input tokensfresh prompt
800Output tokensmodel reply
180Cache read tokensre-read context
50Kcost / request $0.0020 · in $0.14 · out $0.58 · cache $0.04 per M
fresh input 6%output 5%cache reads 89%
Requests / 30 days
5.1K
$10 credits ÷ $0.0020 per request
~1.0K quick fixes~203 bug fixes~34 feature PRs
what real work costs
Real coding tasks priced end to end on Tencent Hy3, from a quick lookup to a full-repo agent run.
One agent task
$0.02
180K in at 75% cache hit, 12K out
What you pay
$0.09 /M
all-in across every token that task touched
Sticker input
$0.14 /M
cache reads bill at $0.04 /M instead
| Task | Tokens in · out | Tencent Hy3 | Muse Spark 1.2 Contributor | Claude Haiku 4.5 |
|---|---|---|---|---|
| Quick lookup / one-liner | 8K · 1K | $0.0009 | $0.0003 | $0.0065 |
| Review a 500-line PR | 60K · 4K | $0.0063 | $0.0027 | $0.04 |
| Fix a bug (agent loop) | 180K · 12K | $0.02 | $0.0072 | $0.12 |
| Refactor a module | 320K · 20K | $0.03 | $0.01 | $0.21 |
| Full-repo agent run | 900K · 45K | $0.08 | $0.03 | $0.49 |
frequently asked
How much does Tencent Hy3 cost?
$0.14/M input and $0.58/M output, cache reads $0.04/M. In an agent loop most input is cache-read, so the effective input rate is about $0.07/M.Which plan do I need?
Available on Go and above.
How do I switch to it?
Run
cmd --model tencent/hy3-paid, 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/tencent-hy3