Model comparison

Muse Spark 1.3 vs Qwen 3.8 Max

Meta against Qwen, compared on context, price, and attributed benchmark results.

Recorded profile verification October 2, 2026Individual provider fields may changeEvidence confidence: Medium — see receipts below

Meta

Muse Spark 1.3

Frontier

Qwen

Qwen 3.8 Max

Frontier

AI model capability comparison
Specification	Muse Spark 1.3	Qwen 3.8 Max
Provider
Provider	Meta	Qwen
Tier
Tier	Frontier	Frontier
Context window
Context window	Winner: 1.05M	1M
Max output
Max output	Not verifiedUnverified	131K
Input / 1M tokens
Input / 1M tokens	Winner: $1.25	$2
Output / 1M tokens
Output / 1M tokens	Winner: $4.25	$6
Weights
Weights	Closed	Closed
Parameters
Parameters	Not verifiedUnverified	2.4T total (MoE; managed active count not verified)
Reasoning levels
Reasoning levels	minimal, low, medium, high, xhigh, max	low, medium, xhigh
Modalities
Modalities	text, image, video, audio, pdf	text, image, video
API model id
API model id	muse-spark-1.3	qwen3.8-max
Released
Released	September 2, 2026	August 3, 2026
Artificial Analysis Intelligence Index [max] · source date Muse Spark 1.3: 2026-10-01
Artificial Analysis Intelligence Index [max] · source date Muse Spark 1.3: 2026-10-01	48	Not verifiedUnverified
Artificial Analysis Intelligence Index · source date Qwen 3.8 Max: 2026-10-01
Artificial Analysis Intelligence Index · source date Qwen 3.8 Max: 2026-10-01	Not verifiedUnverified	45
DeepSWE 1.1 · source date Muse Spark 1.3: 2026-09-02
DeepSWE 1.1 · source date Muse Spark 1.3: 2026-09-02	75.4	Not verifiedUnverified
Terminal-Bench 2.1 · source date Qwen 3.8 Max: 2026-10-01
Terminal-Bench 2.1 · source date Qwen 3.8 Max: 2026-10-01	Not verifiedUnverified	86.6
GPQA Diamond · source date Qwen 3.8 Max: 2026-10-01
GPQA Diamond · source date Qwen 3.8 Max: 2026-10-01	Not verifiedUnverified	92.6
Humanity's Last Exam · source date Qwen 3.8 Max: 2026-08-14
Humanity's Last Exam · source date Qwen 3.8 Max: 2026-08-14	Not verifiedUnverified	56.2

Dates beside benchmark scores refer to the linked source’s publication or capture date unless that source explicitly states when it measured the result. A shared URL alone does not establish comparable runs; unverified setups are marked. Prices are USD per million tokens at standard rates, excluding batch and caching discounts. Values we could not confirm from the provider are shown as “Not verified” rather than estimated.

Benchmark receipts

Where each score comes from, and how far it can be compared across models.

Source date2026-10-01
Qwen official checkpoint card reproducing the managed Max evaluation table (vendor) (Managed Qwen3.8-Max column; complete run configuration unverified)
Graduate-level science reasoning: multiple-choice questions written by domain PhDs.
Comparable with caveatDo not treat a fixed two-point band as universal measurement uncertainty; it depends on run count and evaluation setup. Chain-of-thought, prompt, and tool policy must match. The small pool limits fine-grained conclusions, and frontier scores can be near ceiling; Artificial Analysis dropped it from its Intelligence Index in v4.3.
Source date2026-10-01
Qwen official checkpoint card reproducing the managed Max evaluation table (vendor) (Managed Qwen3.8-Max column; Claude Code avg@10, 5-hour timeout, max_tokens 131,072; exact run date unverified)
Agentic terminal work: multi-step tasks executed in a sandboxed shell environment.
Comparable with caveatNot comparable with Terminal-Bench 3.0 or 4.0 (different task sets) or v1; harness configuration (container, time limits) also shifts results.
Source date2026-10-01
Artificial Analysis (Qwen3.8 Max (0902))
Composite index blending reasoning, knowledge, and coding evals into one 0–100 score.
Comparable with caveatOnly same-version scores are comparable. v4.3.2 was rebaselined (the top score fell from 66 on v4.1.1 to ~58), so v4.1.1 figures must not be compared with v4.3.2 figures — check the cited source and its date.
Source date2026-10-01
Artificial Analysis (Muse Spark 1.3 (max))
Composite index blending reasoning, knowledge, and coding evals into one 0–100 score.
Comparable with caveatOnly same-version scores are comparable. v4.3.2 was rebaselined (the top score fell from 66 on v4.1.1 to ~58), so v4.1.1 figures must not be compared with v4.3.2 figures — check the cited source and its date.
Source date2026-09-02
Meta launch table (Meta max / mini-swe-agent; vendor chart 75.4 versus Spark 1.2 xhigh 55.0; source publication September 2, experiment date unverified; separate from Datacurve xhigh row)
Agentic software engineering: pass@1 on real repo tasks with reported cost per completed task.
Directly comparable
Source date2026-08-14
Z.ai GLM-5.3 vendor evaluation of Qwen3.8-Max, with tools; distinct from Qwen no-tools results
Frontier-knowledge ceiling: extremely hard multi-domain questions written to be near-impossible without deep expertise.
Comparable with caveatSubset/tool configurations (text-only vs with-tools) differ between vendors and materially change scores; must match configuration to compare.

Pricing tiers

Muse Spark 1.3: Meta Standard $1.25/$4.25 per MTok, cached $0.15; prompts/completions are not used for training. Contributor uses a distinct SKU at $0.10/$0.20, cached $0.002, and permits training on prompts/completions. No long-context premium. Search grounding adds $2.50 per 1,000 queries.

Qwen 3.8 Max: Model Studio International $2/$6 per MTok; implicit cache $0.25, explicit creation $2.50/read $0.17. China/Global $1.65/$4.951, cache $0.206, creation $2.063/read $0.137. Output 131,072 in both modes; input 991,808 non-thinking or 983,616 thinking, 1M context and 262,144 reasoning. Batch supported in China, unavailable in International/Global.

FrontierRecord checked October 2, 2026

Muse Spark 1.3

Muse Spark 1.3 is Meta's September 2026 frontier refresh with a 1M-token context for coding and agent workflows.

Best for

Long-horizon coding agents
Frontier quality below frontier pricing
Meta ecosystem

Watch out

Max effort is Standard Spark 1.3 only; Contributor and Spark 1.2 lack it. None returns HTTP 400; omitted effort is model-determined. Reasoning counts toward output budget and billing. Spark 1.3 audio may degrade; Meta directs audio work to 1.2 or Voice Transcribe. Standard 3,000 RPM/4M TPM versus Contributor 100 RPM/3M TPM quotas apply per team. Exact output cap and parameter count remain unknown. Meta's 75.4% DeepSWE max/mini-swe-agent report is separate from independent xhigh observations.

FrontierRecord checked October 2, 2026

Qwen 3.8 Max

Qwen 3.8 Max is Alibaba's managed multimodal flagship, with a vendor-described 2.4T MoE scale and 1M context.

Best for

Hosted agentic work
Long-context
Multimodal

Watch out

The text-only open checkpoint has 262,144 native context with extension and requires thinking; its 95B active count is not verified for the managed API. Managed Max accepts visual input and supports thinking on/off, xhigh default and low/medium/xhigh controls, with preserved thinking by default. API reasoning_effort and thinking_budget cannot both be set. Regional web-search and Batch support differ. Independent AA 0902 results identify a snapshot, not every alias version.

When the cheaper one wins

Muse Spark 1.3 is cheaper on output at $4.25 per million tokens against $6 for Qwen 3.8 Max — about 1.4×. Use the cheaper tier for classification, extraction, summarisation, and any task where the expensive model’s extra score does not change the accepted output. The expensive one only pays if your hardest task actually fails on the cheap tier. This compares standard short-context token rates, not total cost per accepted task. Include input, cache, context tiers, schedules, retries and review in your pilot.

Run the model picker

Evidence confidence: Medium

How strong and complete the evidence behind this comparison is — not a prediction of which model is better.

Pricing verified on both sides — Input and output rates are verified for both models.
3/5 core specs verified on both sides — Not published for at least one side: max output, parameter count.
No directly aligned benchmark setup — No benchmark has matching cited sources, source dates, units, effort and recorded run details on both models.
Verified within the last 90 days — Older of the two catalog checks was 24 days ago.
Both models carry source citations — Each side has at least two catalog sources on record.

Source receipts

Each catalog figure was checked against the provider or an independent second source on the date shown.

Muse Spark 1.3: Meta — Models (accessed 2026-10-01)
Muse Spark 1.3: Meta — Pricing and rate limits (accessed 2026-10-01)
Muse Spark 1.3: Meta — Reasoning (accessed 2026-10-01)
Muse Spark 1.3: Meta — Chat completions schemas (accessed 2026-10-01)
Muse Spark 1.3: Meta — Introducing Muse Spark 1.3 (accessed 2026-10-01)
Muse Spark 1.3: Meta — Muse Spark 1.3 Multimodal Evaluation Methodology (accessed 2026-10-01)
Muse Spark 1.3: Meta — Meta vendor benchmark scorecard (visually checked) (accessed 2026-10-01)
Qwen 3.8 Max: Dedicated exact capacities and regional pricing (accessed 2026-10-01)
Qwen 3.8 Max: Exact effort and compatibility rules (accessed 2026-10-01)
Qwen 3.8 Max: Current managed card and exact cache rates (accessed 2026-10-01)
Qwen 3.8 Max: Current hosted thinking versus checkpoint mode (accessed 2026-10-01)
Qwen 3.8 Max: Qwen — Qwen 3.8 Max (accessed 2026-08-29)
Qwen 3.8 Max: Alibaba Cloud Model Studio (accessed 2026-08-29)
Qwen 3.8 Max: Alibaba Cloud Model Studio — text generation models (1M ctx, 131,072 out) (accessed 2026-09-26)
Qwen 3.8 Max: Qwen3.8-2.4T-A95B checkpoint card (distinguishes enhanced managed Max) (accessed 2026-10-01)
Qwen 3.8 Max: QwenCloud managed Qwen3.8-Max model card (cache and mode-specific limits) (accessed 2026-10-01)

Related comparisons

Diving deeper on one model? Muse Spark 1.3 · Qwen 3.8 Max

Common questions

Muse Spark 1.3 vs Qwen 3.8 Max

Answered from the verified figures on this page rather than general guidance.

Is Muse Spark 1.3 or Qwen 3.8 Max cheaper for input?

Muse Spark 1.3 is cheaper at $1.25 per million input tokens, against $2 for Qwen 3.8 Max — roughly 1.6× the price. Estimate both input and output token volume for your workload. These are base rates; Muse Spark 1.3: Meta Standard $1.25/$4.25 per MTok, cached $0.15; prompts/completions are not used for training. Contributor uses a distinct SKU at $0.10/$0.20, cached $0.002, and permits training on prompts/completions. No long-context premium. Search grounding adds $2.50 per 1,000 queries. Qwen 3.8 Max: Model Studio International $2/$6 per MTok; implicit cache $0.25, explicit creation $2.50/read $0.17. China/Global $1.65/$4.951, cache $0.206, creation $2.063/read $0.137. Output 131,072 in both modes; input 991,808 non-thinking or 983,616 thinking, 1M context and 262,144 reasoning. Batch supported in China, unavailable in International/Global.

Is Muse Spark 1.3 or Qwen 3.8 Max cheaper for output?

Muse Spark 1.3 is cheaper at $4.25 per million output tokens, against $6 for Qwen 3.8 Max — roughly 1.4× the price. Estimate both input and output token volume for your workload. These are base rates; Muse Spark 1.3: Meta Standard $1.25/$4.25 per MTok, cached $0.15; prompts/completions are not used for training. Contributor uses a distinct SKU at $0.10/$0.20, cached $0.002, and permits training on prompts/completions. No long-context premium. Search grounding adds $2.50 per 1,000 queries. Qwen 3.8 Max: Model Studio International $2/$6 per MTok; implicit cache $0.25, explicit creation $2.50/read $0.17. China/Global $1.65/$4.951, cache $0.206, creation $2.063/read $0.137. Output 131,072 in both modes; input 991,808 non-thinking or 983,616 thinking, 1M context and 262,144 reasoning. Batch supported in China, unavailable in International/Global.

Which has the larger context window, Muse Spark 1.3 or Qwen 3.8 Max?

Muse Spark 1.3 accepts 1.05M tokens against 1M for Qwen 3.8 Max. This only matters if you routinely send very long documents or large codebases.

Do Muse Spark 1.3 and Qwen 3.8 Max support the same reasoning levels?

Muse Spark 1.3 exposes minimal, low, medium, high, xhigh, max, while Qwen 3.8 Max exposes low, medium, xhigh.

Should I use Muse Spark 1.3 or Qwen 3.8 Max?

Both are classified in our frontier tier; this editorial grouping does not imply equal capability. Muse Spark 1.3 is a candidate for long-horizon coding agents; Qwen 3.8 Max is a candidate for hosted agentic work. Compare accuracy, tool reliability, latency and cost on representative tasks.

Next step

Choosing between them

The verified figures that separate this pair, computed from the catalog rather than restated boilerplate.

Input price: Muse Spark 1.3 costs $1.25 per 1M tokens versus $2 for Qwen 3.8 Max — a 1.6x difference at the headline tier.
Context: Muse Spark 1.3 takes 1.05M against 1M for Qwen 3.8 Max — only decisive if your prompts approach the smaller window.

Run the model picker Estimate stack cost All model comparisons

Choosing a harness rather than a model — Cursor, Copilot, Claude Code, Windsurf, or Lovable? Compare agentic harnesses · Latest releases.