Model comparison
Claude Sonnet 5 vs Qwen3-235B-A22B
Anthropic against Qwen, compared on context, price, and verified benchmark results.
Catalog record checked September 26, 2026Individual provider fields may changeEvidence confidence: High — see receipts below
Anthropic
Claude Sonnet 5
Balanced
Qwen
Qwen3-235B-A22B
Balanced · Open weights
| Specification | Claude Sonnet 5 | Qwen3-235B-A22B |
|---|---|---|
| Provider | ||
| Provider | Anthropic | Qwen |
| Tier | ||
| Tier | Balanced | Balanced |
| Context window | ||
| Context window | Winner: 1M | 128K |
| Max output | ||
| Max output | Winner: 128K | 33K |
| Input / 1M tokens | ||
| Input / 1M tokens | $2 | Winner: $0.70 |
| Output / 1M tokens | ||
| Output / 1M tokens | $10 | Winner: $2.80 |
| Weights | ||
| Weights | Closed | Open |
| Parameters | ||
| Parameters | Not disclosedUnverified | 235B total / 22B active (MoE) |
| Reasoning levels | ||
| Reasoning levels | adaptive | low, high, max |
| Modalities | ||
| Modalities | text, image | text |
| License | ||
| License | Not disclosedUnverified | Apache 2.0 |
| API model id | ||
| API model id | claude-sonnet-5 | qwen3-235b-a22b |
| Released | ||
| Released | June 30, 2026 | April 29, 2025 |
| Artificial Analysis Intelligence Index [max] (2026-09-26) | ||
| Artificial Analysis Intelligence Index [max] (2026-09-26) | 38.2 | Not verifiedUnverified |
| Artificial Analysis Intelligence Index (2026-09-26) | ||
| Artificial Analysis Intelligence Index (2026-09-26) | Not verifiedUnverified | 9.5 |
| Terminal-Bench 2.1 (2026-06-30) | ||
| Terminal-Bench 2.1 (2026-06-30) | 80.4 | Not verifiedUnverified |
| Humanity's Last Exam (2026-06-30) | ||
| Humanity's Last Exam (2026-06-30) | 57.4 | Not verifiedUnverified |
| SWE-bench Verified (2026-09-01) | ||
| SWE-bench Verified (2026-09-01) | 79.6 | Not verifiedUnverified |
| LiveBench (2026-09-05) | ||
| LiveBench (2026-09-05) | 76 | Not verifiedUnverified |
Prices are USD per million tokens at standard rates, excluding batch and caching discounts. Bold indicates the better figure where one is objectively better. 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.
- 2026-09-26Artificial Analysis (Reasoning, AA-estimated)
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 measuredAt.
- 2026-09-05LiveBench official leaderboard (benchmark-owned), xhigh effort
Contamination-resistant general capability across reasoning, coding, math, data analysis, and language, with monthly question refreshes.
Comparable with caveatRolling question set: observations months apart measure different question mixes. Record the measurement date and compare within ~1 month windows.
- 2026-09-01vals.ai SWE-bench Verified leaderboard (independent, mini-swe-agent bash-only); not published by Anthropic; a 72.7 CosmicJS claim is unverified
Real GitHub issue resolution: does the model's patch pass the hidden tests.
Comparable with caveatPost-audit vendor claims and pre-audit scores sit on different task trust levels; scaffolding (agent harness, compute budget) also dominates results. Never aggregate across scaffolds.
- 2026-06-30Anthropic announcement (vendor, with tools)
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.
- 2026-06-30Anthropic announcement table (vendor, default effort; via Vellum transcription)
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.
Pricing tiers
Claude Sonnet 5: Standard $2/$10 per MTok; cache reads 10% of base input. Knowledge cutoff Jan 2026.
Qwen3-235B-A22B: Model Studio International $0.70 input / $2.80 output per MTok (non-thinking); thinking-mode output $8.40. Open weights (Apache 2.0) for self-hosting; 32,768-token recommended output.
Claude Sonnet 5
Claude Sonnet 5 is Anthropic's best speed/intelligence balance for high-volume agentic work at $2/$10.
Best for
- Everyday generation
- High-volume agents
- Cost-sensitive work
Watch out
A step below Opus 5.5 on the hardest tasks; pick Opus 5.5 when quality is paramount. Sonnet 5.5 is announced but not yet released.
Qwen3-235B-A22B
Qwen3-235B-A22B is Alibaba's widely deployed open-weight MoE — 235B/22B, strong price-performance.
Best for
- Open-weight deployments
- Self-hosting
- Multilingual
Watch out
Needs multi-GPU for the 235B total; hosted rates vary.
When the cheaper one wins
Qwen3-235B-A22B is cheaper on output at $2.80 per million tokens against $10 for Claude Sonnet 5 — about 3.6×. 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. These are standard-tier API rates, excluding batch and cache discounts.
Run the model pickerEvidence confidence: High
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.
- 1 shared named benchmark with differing scores — Measured on: Artificial Analysis Intelligence Index.
- Verified within the last 90 days — Newest catalog check was 2 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 Sonnet 5: Anthropic — Claude Sonnet 5 (accessed 2026-08-29)
- Claude Sonnet 5: Anthropic news — Claude Sonnet 5 (accessed 2026-08-29)
- Qwen3-235B-A22B: HuggingFace — Qwen3-235B-A22B (accessed 2026-08-29)
- Qwen3-235B-A22B: Qwen — Qwen3 (accessed 2026-08-29)
- Qwen3-235B-A22B: Alibaba Cloud Model Studio pricing (qwen3-235b-a22b $0.70/$2.80, thinking $8.40) (accessed 2026-09-26)
Related comparisons
- Amazon Nova 2 Pro vs Claude Sonnet 5
- Amazon Nova 2 Pro vs Qwen3-235B-A22B
- Claude Fable 5.1 vs Claude Sonnet 5
- Claude Fable 5 vs Claude Sonnet 5
- Claude Haiku 4.5 vs Claude Sonnet 5
- Claude Mythos 5.1 vs Claude Sonnet 5
Diving deeper on one model? Claude Sonnet 5 · Qwen3-235B-A22B
Common questions
Claude Sonnet 5 vs Qwen3-235B-A22B
Answered from the verified figures on this page rather than general guidance.
Is Claude Sonnet 5 or Qwen3-235B-A22B cheaper for input?
Is Claude Sonnet 5 or Qwen3-235B-A22B cheaper for output?
Which has the larger context window, Claude Sonnet 5 or Qwen3-235B-A22B?
Do Claude Sonnet 5 and Qwen3-235B-A22B support the same reasoning levels?
Should I use Claude Sonnet 5 or Qwen3-235B-A22B?
Can I self-host Claude Sonnet 5 or Qwen3-235B-A22B?
Next step
Choosing between them
The verified figures that separate this pair, computed from the catalog rather than restated boilerplate.
- Input price: Qwen3-235B-A22B costs $0.70 per 1M tokens versus $2 for Claude Sonnet 5 — a 2.9x difference at the headline tier.
- Context: Claude Sonnet 5 takes 1M against 128K for Qwen3-235B-A22B — only decisive if your prompts approach the smaller window.
- Deployment: Qwen3-235B-A22B publishes weights you can self-host; the other is API-only.
Choosing a harness rather than a model — Cursor, Copilot, Claude Code, Windsurf, or Lovable? Compare agentic harnesses · Latest releases.