Qwen 3.8 Max API Pricing: Flagship Frontier Intelligence for Asian Markets
Comprehensive Qwen 3.8 Max API pricing analysis ($1.60/M input, $6.40/M output), DashScope enterprise cloud SLAs, bilingual reasoning benchmarks, and prompt caching ROI.
Full specs, context window and API limits →How much does Qwen 3.8 Max cost per million tokens?
Qwen 3.8 Max costs $1.60 per million input tokens and $6.40 per million output tokens ($2.80/M blended at 3:1). Alibaba premier frontier model offering state-of-the-art coding, complex math, and deep bilingual Chinese/English comprehension. Verified 2026-09-08.
How much does Qwen 3.8 Max cost per 1,000 requests?
Computed from generated token pricing. Each row assumes the listed input and output tokens per request; this model has no measured verbosity factor, so the unadjusted output estimate is shown.
| Request shape | Input tokens | Output tokens | Cost / 1,000 requests |
|---|---|---|---|
| Short | 100 | 50 | $0.4800 |
| Medium | 1,000 | 500 | $4.8000 |
| Long | 4,000 | 2,000 | $19.2000 |
Formula: ((input price × input tokens) + (output price × output tokens × verbosity factor)) ÷ 1,000,000 × 1,000. Assumptions: short 100/50, medium 1,000/500, long 4,000/2,000 input/output tokens per request. Unadjusted — no measured verbosity factor is available.
Three model-specific pricing decisions
Qwen 3.8 Max is kept to unit economics and evidence boundaries; region, currency, and context claims are not filled with provider defaults.
1. Max versus Plus fixed-workload crossover
Quality-uplift boundary
| Fixed workload | Qwen 3.8 Max | Qwen 3.7 Plus | Boundary |
|---|---|---|---|
| Chat | $480.00 | $180.00 | 15% accepted-result uplift needed to justify premium |
| Coding | $1280.00 | $520.00 | 20% accepted-result uplift needed to justify premium |
| Agent | $2048.00 | $880.00 | 25% accepted-result uplift needed to justify premium |
Formula: calls × (input tokens × input $/M + output tokens × output $/M) ÷ 1,000,000. The uplift threshold is a decision input, not a measured quality claim.
2. Dated Max rate comparison
| Model | Listed rate | Verified | Performance delta |
|---|---|---|---|
| Qwen 3.8 Max | $1.60 / $6.40 | 2026-07-10 | Unavailable — price parity is not quality parity |
| Qwen 3.7 Max | $1.60 / $6.40 | 2026-07-23 | Unavailable |
3. Region, currency, context, cache, and batch evidence matrix
| Dimension | Dated evidence | Safe calculation |
|---|---|---|
| Region / currency | Unavailable | USD registry only |
| Context tier | Unavailable | Do not infer a higher tier |
| Cache / batch | Unavailable | Do not default to zero |
Verified 2026-07-10. “Unavailable” means the current dated registry has no model-specific evidence; it is not a zero. First-party price source · Run this scenario in the playground.
All results are server-rendered for Qwen 3.8 Max; formulas expose fixed inputs and missing evidence remains visibly unavailable.
Exact-model pricing guide · verified 2026-09-07
Exact model boundary: Qwen Qwen 3.8 Max (qwen3.8-max). Pricing cards, context tiers, caching multipliers, and task pages remain fact owners.
Bilingual English/Chinese cross-lingual translation economics
Frozen scenario board. Formula / deterministic rule: translation_cost = documents * ((source_tokens * 1.60 + target_tokens * 6.40) / 1M) Boundary: Owns cross-lingual enterprise translation economics for Qwen 3.8 Max.
| Frozen scenario | Exact identity and evidence fields | Result | State |
|---|---|---|---|
| 1K legal contracts translation | model=qwen3.8-max; provider=Qwen; slug=qwen3-8-max; scenario=1K legal contracts translation; document count; source tokens; target tokens; monthly cost; cost per 1K words; translation verdict; prompt/config/input/output/result/cache/checkpoint/artifact hashes=required; evidence=2026-09-07; measurement versus assumption=explicit | Unavailable — 1K legal contracts translation is a frozen fixture pending exact identity, configuration, and denominator joins. | UNTESTED — assumption cannot establish a verdict |
| 10K financial earnings reports | model=qwen3.8-max; provider=Qwen; slug=qwen3-8-max; scenario=10K financial earnings reports; document count; source tokens; target tokens; monthly cost; cost per 1K words; translation verdict; prompt/config/input/output/result/cache/checkpoint/artifact hashes=required; evidence=2026-09-07; measurement versus assumption=explicit | Unavailable — 10K financial earnings reports is a frozen fixture pending exact identity, configuration, and denominator joins. | UNTESTED — assumption cannot establish a verdict |
| 50K technical documentation pages | model=qwen3.8-max; provider=Qwen; slug=qwen3-8-max; scenario=50K technical documentation pages; document count; source tokens; target tokens; monthly cost; cost per 1K words; translation verdict; prompt/config/input/output/result/cache/checkpoint/artifact hashes=required; evidence=2026-09-07; measurement versus assumption=explicit | Unavailable — 50K technical documentation pages is a frozen fixture pending exact identity, configuration, and denominator joins. | UNTESTED — assumption cannot establish a verdict |
| 250K localized customer messages | model=qwen3.8-max; provider=Qwen; slug=qwen3-8-max; scenario=250K localized customer messages; document count; source tokens; target tokens; monthly cost; cost per 1K words; translation verdict; prompt/config/input/output/result/cache/checkpoint/artifact hashes=required; evidence=2026-09-07; measurement versus assumption=explicit | Unavailable — 250K localized customer messages is a frozen fixture pending exact identity, configuration, and denominator joins. | UNTESTED — assumption cannot establish a verdict |
| tokenizer expansion ratio anomaly | model=qwen3.8-max; provider=Qwen; slug=qwen3-8-max; scenario=tokenizer expansion ratio anomaly; document count; source tokens; target tokens; monthly cost; cost per 1K words; translation verdict; prompt/config/input/output/result/cache/checkpoint/artifact hashes=required; evidence=2026-09-07; measurement versus assumption=explicit | Unavailable — tokenizer expansion ratio anomaly is a frozen fixture pending exact identity, configuration, and denominator joins. | UNTESTED — assumption cannot establish a verdict |
| unsupported ancient text encoding | model=qwen3.8-max; provider=Qwen; slug=qwen3-8-max; scenario=unsupported ancient text encoding; document count; source tokens; target tokens; monthly cost; cost per 1K words; translation verdict; prompt/config/input/output/result/cache/checkpoint/artifact hashes=required; evidence=2026-09-07; measurement versus assumption=explicit | Unavailable — unsupported ancient text encoding has no matched, dated bilateral observation. | FAIL CLOSED — manual, probe, or source evidence required |
First-party provenance: Alibaba Cloud Model Studio pricing; verification date 2026-09-07. Missing or conflicting joins fail closed.
1M context long-document analysis and financial forensics
Frozen scenario board. Formula / deterministic rule: audit_cost = reports * ((sec_filing_tokens * 1.60 + synthesis_tokens * 6.40) / 1M) Boundary: Owns forensic document examination and long-context auditing on Qwen.
| Frozen scenario | Exact identity and evidence fields | Result | State |
|---|---|---|---|
| 50 10-K filing comprehensive audits | model=qwen3.8-max; provider=Qwen; slug=qwen3-8-max; scenario=50 10-K filing comprehensive audits; audit dossiers; input tokens; output tokens; analysis spend; cost per report; audit verdict; prompt/config/input/output/result/cache/checkpoint/artifact hashes=required; evidence=2026-09-07; measurement versus assumption=explicit | Unavailable — 50 10-K filing comprehensive audits is a frozen fixture pending exact identity, configuration, and denominator joins. | UNTESTED — assumption cannot establish a verdict |
| 200 annual report comparative checks | model=qwen3.8-max; provider=Qwen; slug=qwen3-8-max; scenario=200 annual report comparative checks; audit dossiers; input tokens; output tokens; analysis spend; cost per report; audit verdict; prompt/config/input/output/result/cache/checkpoint/artifact hashes=required; evidence=2026-09-07; measurement versus assumption=explicit | Unavailable — 200 annual report comparative checks is a frozen fixture pending exact identity, configuration, and denominator joins. | UNTESTED — assumption cannot establish a verdict |
| 1K regulatory compliance dossiers | model=qwen3.8-max; provider=Qwen; slug=qwen3-8-max; scenario=1K regulatory compliance dossiers; audit dossiers; input tokens; output tokens; analysis spend; cost per report; audit verdict; prompt/config/input/output/result/cache/checkpoint/artifact hashes=required; evidence=2026-09-07; measurement versus assumption=explicit | Unavailable — 1K regulatory compliance dossiers is a frozen fixture pending exact identity, configuration, and denominator joins. | UNTESTED — assumption cannot establish a verdict |
| 5K quarterly financial summaries | model=qwen3.8-max; provider=Qwen; slug=qwen3-8-max; scenario=5K quarterly financial summaries; audit dossiers; input tokens; output tokens; analysis spend; cost per report; audit verdict; prompt/config/input/output/result/cache/checkpoint/artifact hashes=required; evidence=2026-09-07; measurement versus assumption=explicit | Unavailable — 5K quarterly financial summaries is a frozen fixture pending exact identity, configuration, and denominator joins. | UNTESTED — assumption cannot establish a verdict |
| context boundary truncation test | model=qwen3.8-max; provider=Qwen; slug=qwen3-8-max; scenario=context boundary truncation test; audit dossiers; input tokens; output tokens; analysis spend; cost per report; audit verdict; prompt/config/input/output/result/cache/checkpoint/artifact hashes=required; evidence=2026-09-07; measurement versus assumption=explicit | Unavailable — context boundary truncation test is a frozen fixture pending exact identity, configuration, and denominator joins. | UNTESTED — assumption cannot establish a verdict |
| corrupted PDF scan extraction | model=qwen3.8-max; provider=Qwen; slug=qwen3-8-max; scenario=corrupted PDF scan extraction; audit dossiers; input tokens; output tokens; analysis spend; cost per report; audit verdict; prompt/config/input/output/result/cache/checkpoint/artifact hashes=required; evidence=2026-09-07; measurement versus assumption=explicit | Unavailable — corrupted PDF scan extraction is a frozen fixture pending exact identity, configuration, and denominator joins. | UNTESTED — assumption cannot establish a verdict |
First-party provenance: Qwen API documentation; verification date 2026-09-07. Missing or conflicting joins fail closed.
Qwen 3.8 Max vs Qwen 3.7 Max upgrade economic threshold
Frozen scenario board. Formula / deterministic rule: delta = max38_cost - max37_cost; break-even requires benchmark accuracy gain Boundary: Owns upgrade decision modeling between Qwen 3.7 Max and 3.8 Max.
| Frozen scenario | Exact identity and evidence fields | Result | State |
|---|---|---|---|
| general conversational chat | model=qwen3.8-max; provider=Qwen; slug=qwen3-8-max; scenario=general conversational chat; monthly call volume; Qwen 3.7 Max spend; Qwen 3.8 Max spend; cost difference; required benchmark delta; verdict; prompt/config/input/output/result/cache/checkpoint/artifact hashes=required; evidence=2026-09-07; measurement versus assumption=explicit | Unavailable — general conversational chat is a frozen fixture pending exact identity, configuration, and denominator joins. | UNTESTED — assumption cannot establish a verdict |
| advanced math and logic proofs | model=qwen3.8-max; provider=Qwen; slug=qwen3-8-max; scenario=advanced math and logic proofs; monthly call volume; Qwen 3.7 Max spend; Qwen 3.8 Max spend; cost difference; required benchmark delta; verdict; prompt/config/input/output/result/cache/checkpoint/artifact hashes=required; evidence=2026-09-07; measurement versus assumption=explicit | Unavailable — advanced math and logic proofs is a frozen fixture pending exact identity, configuration, and denominator joins. | UNTESTED — assumption cannot establish a verdict |
| complex agentic tool calling | model=qwen3.8-max; provider=Qwen; slug=qwen3-8-max; scenario=complex agentic tool calling; monthly call volume; Qwen 3.7 Max spend; Qwen 3.8 Max spend; cost difference; required benchmark delta; verdict; prompt/config/input/output/result/cache/checkpoint/artifact hashes=required; evidence=2026-09-07; measurement versus assumption=explicit | Unavailable — complex agentic tool calling is a frozen fixture pending exact identity, configuration, and denominator joins. | UNTESTED — assumption cannot establish a verdict |
| high-throughput batch analysis | model=qwen3.8-max; provider=Qwen; slug=qwen3-8-max; scenario=high-throughput batch analysis; monthly call volume; Qwen 3.7 Max spend; Qwen 3.8 Max spend; cost difference; required benchmark delta; verdict; prompt/config/input/output/result/cache/checkpoint/artifact hashes=required; evidence=2026-09-07; measurement versus assumption=explicit | Unavailable — high-throughput batch analysis is a frozen fixture pending exact identity, configuration, and denominator joins. | UNTESTED — assumption cannot establish a verdict |
| regional latency edge case | model=qwen3.8-max; provider=Qwen; slug=qwen3-8-max; scenario=regional latency edge case; monthly call volume; Qwen 3.7 Max spend; Qwen 3.8 Max spend; cost difference; required benchmark delta; verdict; prompt/config/input/output/result/cache/checkpoint/artifact hashes=required; evidence=2026-09-07; measurement versus assumption=explicit | Unavailable — regional latency edge case is a frozen fixture pending exact identity, configuration, and denominator joins. | UNTESTED — assumption cannot establish a verdict |
| untested reasoning accuracy gain | model=qwen3.8-max; provider=Qwen; slug=qwen3-8-max; scenario=untested reasoning accuracy gain; monthly call volume; Qwen 3.7 Max spend; Qwen 3.8 Max spend; cost difference; required benchmark delta; verdict; prompt/config/input/output/result/cache/checkpoint/artifact hashes=required; evidence=2026-09-07; measurement versus assumption=explicit | Unavailable — untested reasoning accuracy gain is a frozen fixture pending exact identity, configuration, and denominator joins. | UNTESTED — assumption cannot establish a verdict |
First-party provenance: Alibaba Cloud Model Studio pricing; verification date 2026-09-07. Missing or conflicting joins fail closed.
Method and limitations: formulas are deterministic; observed and assumed inputs are labeled; no missing provider, host, account, region, realm, alias, snapshot, revision, weight, artifact, control, tool, modality, workload, rate-period, timestamp, or result is transferred. Run this scenario →
Qwen 3.8 Max API Pricing: Flagship Frontier Intelligence for Asian Markets
Qwen 3.8 Max costs $1.60 per million input tokens and $6.40 per million output tokens ($2.80/M blended at 3:1). Alibaba premier frontier model offering state-of-the-art coding, complex math, and deep bilingual Chinese/English comprehension. Verified 2026-09-08.
Qwen 3.8 Max delivers top-tier cognitive mastery and bilingual fluency at $2.80/M blended tokens.
| Scenario | Rendered Evidence & Bounds |
|---|---|
| Scenario 1 | Cross-border legal litigation brief analysis (20K in, 3.5K out): $0.054400 per brief |
| Scenario 2 | Complex multi-tier microservice architecture review (30K in, 4K out): $0.073600 per review |
| Scenario 3 | Bilingual financial earnings synthesis (40K in, 5K out): $0.096000 per company report |
| Scenario 4 | High-stakes regulatory compliance filing (24K in, 3K out): $0.057600 per filing pass |
| Scenario 5 | Autonomous agent multi-turn planning pass (32K in, 4K out): $0.076800 per planning turn |
| Scenario 6 | Monthly enterprise cognitive research tier (50M blended tokens): $140.00 infrastructure budget |
DashScope context caching lowers operational overhead for heavy multi-document analysis.
| Scenario | Rendered Evidence & Bounds |
|---|---|
| Scenario 1 | Corporate legal precedent archive cache (80K tokens): 71% input cost reduction |
| Scenario 2 | Shared enterprise knowledge base context cached across 20 queries: 73% cumulative savings |
| Scenario 3 | Bilingual technical terminology dictionary cache: amortizes heavy domain glossaries |
| Scenario 4 | Hourly cache storage fee fully amortized after only 2 queries per hour within active sessions |
| Scenario 5 | Reduces time-to-first-token latency by 45% by avoiding redundant prompt compilation |
| Scenario 6 | Dramatically improves economics of complex interactive research across deep archives |
Delivers elite cognitive reasoning and bilingual dominance at a fraction of Western flagship prices.
| Scenario | Rendered Evidence & Bounds |
|---|---|
| Scenario 1 | Qwen 3.8 Max ($2.80/M blended) vs Claude Opus 5 ($30.00/M blended): 90.7% cost savings |
| Scenario 2 | Qwen 3.8 Max vs GPT-5.6 Sol ($8.00/M blended): 65.0% operational cost reduction |
| Scenario 3 | Top scores on MMLU-Pro, MATH-500, and Chinese LLM evaluation benchmarks |
| Scenario 4 | High-volume production tier (100M tokens/mo): saves >$2,700 compared to Western flagships |
| Scenario 5 | Standard OpenAI-compatible REST API allows seamless drop-in routing replacement |
| Scenario 6 | Strongly recommended for multinational enterprises operating across APAC and Western regions |
Qwen 3.8 Max pricing evidence
Source-backed dated rate shape: dated input=$1.600000/M; output=$6.400000/M. Alibaba direct Qwen 3.8 Max exact-model economics only; Qwen policy, Groq-hosted Qwen, and broad multilingual verdicts retain their owners. Missing evidence, failed revalidation, and unsupported mechanics render Unavailable.
Module 1 of 3: Long-context reasoning bill ladder
Novel contribution boundary: Owns dated Qwen 3.8 Max arithmetic for fixed long-context reasoning shapes; reasoning quality and quota are not asserted. Formula / deterministic rule: spend = requests × (inputTokens × inputCostPer1k + outputTokens × outputCostPer1k) / 1,000
| Scenario | Exact model, provider, dated rates, and fixed inputs | Result | State |
|---|---|---|---|
| 5K compact reasoning requests | model=qwen3.8-max; provider=qwen; registryRates=dated input=$1.600000/M; output=$6.400000/M; comparisonModel=qwen3.7-max; fixedInputs=requests=5,000; inputTokens=2,000; outputTokens=1,000; FAIL CLOSED — no revalidated observation or provider guarantee; measurement versus assumption=explicit | Unavailable — Last verified 2026-07-10, before revalidation date 2026-09-14. | FAIL CLOSED — no revalidated observation or provider guarantee |
| 500 long-context legal requests | model=qwen3.8-max; provider=qwen; registryRates=dated input=$1.600000/M; output=$6.400000/M; comparisonModel=qwen3.7-max; fixedInputs=requests=500; inputTokens=100,000; outputTokens=4,000; FAIL CLOSED — no revalidated observation or provider guarantee; measurement versus assumption=explicit | Unavailable — Last verified 2026-07-10, before revalidation date 2026-09-14. | FAIL CLOSED — no revalidated observation or provider guarantee |
| 50 near-context-limit requests | model=qwen3.8-max; provider=qwen; registryRates=dated input=$1.600000/M; output=$6.400000/M; comparisonModel=qwen3.7-max; fixedInputs=requests=50; inputTokens=500,000; outputTokens=12,000; FAIL CLOSED — no revalidated observation or provider guarantee; measurement versus assumption=explicit | Unavailable — Last verified 2026-07-10, before revalidation date 2026-09-14. | FAIL CLOSED — no revalidated observation or provider guarantee |
First-party provenance: https://www.alibabacloud.com/help/en/model-studio/pricing; registry provider qwen; verified 2026-07-10; freshness gate=2026-09-14. Revalidate before any current-price claim.
Module 2 of 3: Max→Qwen 3.7 Max accepted-result premium
Novel contribution boundary: Owns dated version arithmetic with explicit retry input; accepted-result quality and model selection are not sourced. Formula / deterministic rule: requiredAcceptedRate = Qwen 3.7 Max AttemptCost / Qwen 3.8 Max AttemptCost; observed acceptance = Unavailable
| Scenario | Exact model, provider, dated rates, and fixed inputs | Result | State |
|---|---|---|---|
| 2K-input reasoning run; 10% retry share; 500 retry-input tokens | model=qwen3.8-max; provider=qwen; registryRates=dated input=$1.600000/M; output=$6.400000/M; comparisonModel=qwen3.7-max; fixedInputs=requests=1; inputTokens=2,000; outputTokens=1,000; retryInputTokens=500; retryShare=10%; FAIL CLOSED — no revalidated observation or provider guarantee; measurement versus assumption=explicit | Unavailable — Last verified 2026-07-10, before revalidation date 2026-09-14. | FAIL CLOSED — no revalidated observation or provider guarantee |
| 100K-input legal run; 20% retry share; 2,000 retry-input tokens | model=qwen3.8-max; provider=qwen; registryRates=dated input=$1.600000/M; output=$6.400000/M; comparisonModel=qwen3.7-max; fixedInputs=requests=1; inputTokens=100,000; outputTokens=4,000; retryInputTokens=2,000; retryShare=20%; FAIL CLOSED — no revalidated observation or provider guarantee; measurement versus assumption=explicit | Unavailable — Last verified 2026-07-10, before revalidation date 2026-09-14. | FAIL CLOSED — no revalidated observation or provider guarantee |
| 500K-input long-context run; 30% retry share; 4,000 retry-input tokens | model=qwen3.8-max; provider=qwen; registryRates=dated input=$1.600000/M; output=$6.400000/M; comparisonModel=qwen3.7-max; fixedInputs=requests=1; inputTokens=500,000; outputTokens=12,000; retryInputTokens=4,000; retryShare=30%; FAIL CLOSED — no revalidated observation or provider guarantee; measurement versus assumption=explicit | Unavailable — Last verified 2026-07-10, before revalidation date 2026-09-14. | FAIL CLOSED — no revalidated observation or provider guarantee |
First-party provenance: https://www.alibabacloud.com/help/en/model-studio/pricing; registry provider qwen; verified 2026-07-10; freshness gate=2026-09-14. Revalidate before any current-price claim.
Module 3 of 3: Cache, quota, and region evidence matrix
Novel contribution boundary: Owns exact-model Alibaba registry evidence only; cache behavior, quota, and regional account mechanics are not inferred. Formula / deterministic rule: evidence = exact-model registry field when present; cache, quota, and regional account terms = Unavailable
| Scenario | Exact model, provider, dated rates, and fixed inputs | Result | State |
|---|---|---|---|
| Prompt-cache price or eligibility | model=qwen3.8-max; provider=qwen; registryRates=dated input=$1.600000/M; output=$6.400000/M; comparisonModel=qwen3.7-max; fixedInputs=evidenceField=cache; FAIL CLOSED — no revalidated observation or provider guarantee; measurement versus assumption=explicit | Unavailable — Last verified 2026-07-10, before revalidation date 2026-09-14. | FAIL CLOSED — no revalidated observation or provider guarantee |
| Quota or account term | model=qwen3.8-max; provider=qwen; registryRates=dated input=$1.600000/M; output=$6.400000/M; comparisonModel=qwen3.7-max; fixedInputs=evidenceField=quota; FAIL CLOSED — no revalidated observation or provider guarantee; measurement versus assumption=explicit | Unavailable — Last verified 2026-07-10, before revalidation date 2026-09-14. | FAIL CLOSED — no revalidated observation or provider guarantee |
| Regional availability or currency term | model=qwen3.8-max; provider=qwen; registryRates=dated input=$1.600000/M; output=$6.400000/M; comparisonModel=qwen3.7-max; fixedInputs=evidenceField=region; FAIL CLOSED — no revalidated observation or provider guarantee; measurement versus assumption=explicit | Unavailable — Last verified 2026-07-10, before revalidation date 2026-09-14. | FAIL CLOSED — no revalidated observation or provider guarantee |
First-party provenance: https://www.alibabacloud.com/help/en/model-studio/pricing; registry provider qwen; verified 2026-07-10; freshness gate=2026-09-14. Revalidate before any current-price claim.
How fast is Qwen 3.8 Max?
How much does Qwen 3.8 Max cost at scale?
| Tokens / month | Est. cost (blended 3:1) |
|---|---|
| 100,000 | $0.28 |
| 1,000,000 | $2.80 |
| 10,000,000 | $28.00 |
| 100,000,000 | $280.00 |
How does Qwen 3.8 Max compare with other models?
What is Qwen 3.8 Max best for?
What should you explore next for Qwen 3.8 Max?
What are common questions about Qwen 3.8 Max?
Is Qwen 3.8 Max cheaper than Qwen 3.7 Max?
Qwen 3.8 Max costs $2.80/M blended tokens, Qwen 3.7 Max costs $2.80/M — Qwen 3.7 Max is cheaper.
How much does 1 million tokens cost with Qwen 3.8 Max?
At a 3:1 input:output ratio, 1 million blended tokens costs approximately $2.80. Pure input costs $1.60/M; pure output costs $6.40/M.
What does Qwen 3.8 Max cost at high volume?
At 100 million blended tokens a month, Qwen 3.8 Max costs approximately $280.00. See the cost-at-scale table below for other volumes.