Model comparison

Claude Mythos 5.1 vs Qwen 3.8 Max

Anthropic 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

Anthropic

Claude Mythos 5.1

Frontier

Qwen

Qwen 3.8 Max

Frontier

AI model capability comparison
Specification	Claude Mythos 5.1	Qwen 3.8 Max
Provider
Provider	Anthropic	Qwen
Tier
Tier	Frontier	Frontier
Context window
Context window	1M	1M
Max output
Max output	128K	Winner: 131K
Input / 1M tokens
Input / 1M tokens	$10	Winner: $2
Output / 1M tokens
Output / 1M tokens	$50	Winner: $6
Weights
Weights	Closed	Closed
Parameters
Parameters	Not verifiedUnverified	2.4T total (MoE; managed active count not verified)
Reasoning levels
Reasoning levels	low, medium, high, xhigh, max	low, medium, xhigh
Modalities
Modalities	text, image	text, image, video
API model id
API model id	claude-mythos-5-1	qwen3.8-max
Released
Released	September 1, 2026	August 3, 2026
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
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 date2026-08-14
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

Claude Mythos 5.1: Standard $10/$50 per MTok, sharing Fable 5.1's specifications/pricing. Cache reads $0.25; 5-minute writes $12.50 and 1-hour writes $20. Batch $5/$25. US-only inference adds 10% where supported. Invite-only for vetted US organisations; API ID claude-mythos-5-1. Retirement not sooner than September 1, 2027; this is a minimum commitment, not an announced shutdown.

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 1, 2026

Claude Mythos 5.1

Anthropic describes Claude Mythos 5.1 as its most capable model for cybersecurity and biology research — the Fable 5.1 model with fewer cyber/bio safeguards, restricted to vetted US organisations.

Best for

Cyber-defence research
Biology research
Claude Security workloads

Watch out

Invite-only for vetted US organisations through trusted access programmes, with 30-day retention required by default for safety monitoring. Adaptive thinking cannot be disabled and forced tool use errors. Shared underlying model does not make Fable benchmark results Mythos observations; safeguards and access differ.

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

Qwen 3.8 Max is cheaper on output at $6 per million tokens against $50 for Claude Mythos 5.1 — about 8.3×. 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.
4/5 core specs verified on both sides — Not published for at least one side: 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 25 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.

Claude Mythos 5.1: Claude model overview — whole-profile source check October 1 (accessed 2026-10-01)
Claude Mythos 5.1: Claude current API pricing, caching and inference geography (accessed 2026-10-01)
Claude Mythos 5.1: Anthropic — Claude Mythos (accessed 2026-09-03)
Claude Mythos 5.1: Anthropic — model deprecations (claude-mythos-5) (accessed 2026-09-03)
Claude Mythos 5.1: Anthropic — Mythos 5.1 overview (1M context, 128K output, claude-mythos-5-1) (accessed 2026-09-26)
Claude Mythos 5.1: Anthropic selectable effort levels (separate from adaptive thinking mode) (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? Claude Mythos 5.1 · Qwen 3.8 Max

Common questions

Claude Mythos 5.1 vs Qwen 3.8 Max

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

Is Claude Mythos 5.1 or Qwen 3.8 Max cheaper for input?

Qwen 3.8 Max is cheaper at $2 per million input tokens, against $10 for Claude Mythos 5.1 — roughly 5.0× the price. Estimate both input and output token volume for your workload. These are base rates; Claude Mythos 5.1: Standard $10/$50 per MTok, sharing Fable 5.1's specifications/pricing. Cache reads $0.25; 5-minute writes $12.50 and 1-hour writes $20. Batch $5/$25. US-only inference adds 10% where supported. Invite-only for vetted US organisations; API ID claude-mythos-5-1. Retirement not sooner than September 1, 2027; this is a minimum commitment, not an announced shutdown. 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 Claude Mythos 5.1 or Qwen 3.8 Max cheaper for output?

Qwen 3.8 Max is cheaper at $6 per million output tokens, against $50 for Claude Mythos 5.1 — roughly 8.3× the price. Estimate both input and output token volume for your workload. These are base rates; Claude Mythos 5.1: Standard $10/$50 per MTok, sharing Fable 5.1's specifications/pricing. Cache reads $0.25; 5-minute writes $12.50 and 1-hour writes $20. Batch $5/$25. US-only inference adds 10% where supported. Invite-only for vetted US organisations; API ID claude-mythos-5-1. Retirement not sooner than September 1, 2027; this is a minimum commitment, not an announced shutdown. 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, Claude Mythos 5.1 or Qwen 3.8 Max?

Both accept about 1M tokens of context, so document length will not decide between them.

Do Claude Mythos 5.1 and Qwen 3.8 Max support the same reasoning levels?

Claude Mythos 5.1 exposes low, medium, high, xhigh, max, while Qwen 3.8 Max exposes low, medium, xhigh.

Should I use Claude Mythos 5.1 or Qwen 3.8 Max?

Both are classified in our frontier tier; this editorial grouping does not imply equal capability. Claude Mythos 5.1 is a candidate for cyber-defence research; 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: Qwen 3.8 Max costs $2 per 1M tokens versus $10 for Claude Mythos 5.1 — a 5x difference at the headline tier.

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.