Qwen3.8 2.4T A95B
by Alibaba (Qwen) · qwen/qwen3.8-2.4t-a95b · #318 cheapest of 422 paid models
Prices updated Sep 20, 2026, 12:34 AM UTC · refreshed hourly
Qwen3.8 2.4T A95B is an open-weight sparse mixture-of-experts model from Qwen and the open-weight variant of [Qwen3.8 Max](/qwen/qwen3.8-max), with 95 billion active parameters out of 2.4 trillion total. It is...
Compare vs:Claude Sonnet 5Claude Opus 4.6Claude Haiku 4.5GPT-5.6 Sol
Pricing
Input
$2.00
per 1M tokens
Output
$6.00
per 1M tokens
Blended (3:1)
$3.00
3 input : 1 output
Cache read
$0.25
per 1M cached tokens
Cache write
—
per 1M tokens
Cache & batch economics
Effective prices for Qwen3.8 2.4T A95B with prompt caching.
| Cache hit rate | Effective blended $/1M | Example request* | vs no cache |
|---|---|---|---|
| 0% | $3.00 | $0.21 | — |
| 50% | $2.34 | $0.1225 | −42% |
| 90% | $1.82 | $0.0525 | −75% |
*Example: 100K sticky context + 2K new input + 1K output. Batch discount is an approximate 50% off list rates (available for OpenAI / Anthropic / Google — not flagged for this creator). Confirm on the vendor’s pricing page.
What a request costs
| Workload | Input tokens | Output tokens | Cost / request | Cost / 1K requests |
|---|---|---|---|---|
| Short chat message | 500 | 300 | $0.0028 | $2.80 |
| Document summary | 8,000 | 1,000 | $0.022 | $22.00 |
| Codebase question (RAG) | 30,000 | 2,000 | $0.072 | $72.00 |
| Long-context analysis | 150,000 | 5,000 | $0.33 | $330.00 |
Performance
Intelligence Index
39.9
AA composite quality
Coding Index
71.9
Math Index
—
Agentic Index
50.4
Output speed
39 tok/s
median
Time to first token
1.88s
median
Time to first answer
53.08s
after reasoning tokens
Artificial Analysis benchmarks
| Benchmark | Score |
|---|---|
| GPQA Diamond | 93.5% |
| Humanity's Last Exam | 42.4% |
| SciCode | 54.1% |
| AA-LCR (Long Context Reasoning) | 80.3% |
| Terminal-Bench 2.1 | 82.0% |
| τ³-Bench Banking | 49.1% |
Specs
Context window
1.05M
tokens
Max output
131K
tokens
Input modalities
text
Tokenizer
Qwen