Noometry

Model comparison

Claude Sonnet 4.6 vs Qwen2.5 7B Instruct

Claude Sonnet 4.6 is the stronger model overall, scoring 50.1 to 28.6 on the Noometry Index. Qwen2.5 7B Instruct costs 20× less per token, which makes it the better buy when Claude Sonnet 4.6's lead doesn't matter for your workload.

Last verified . 6 shared benchmarks.

Claude Sonnet 4.6 Anthropic

50.1

Rank #42 Confirmed

Qwen2.5 7B Instruct Alibaba (Qwen)

28.6

Rank #314 Confirmed

Summary

  • They share 6 benchmarks with published results for both. Claude Sonnet 4.6 scores higher in 7 categories and Qwen2.5 7B Instruct in 0 categories; 7 gaps are clear of the uncertainty.
  • The widest gap is in math, where Claude Sonnet 4.6 leads 51.9 to 11.8.
  • The biggest single-benchmark swing is OTIS Mock AIME 2024-2025: 85.8% for Claude Sonnet 4.6 and 2.5% for Qwen2.5 7B Instruct.
  • Qwen2.5 7B Instruct is cheaper at $0.17 / $0.70 per million input/output tokens, against $3 / $15 for Claude Sonnet 4.6.
  • Claude Sonnet 4.6 accepts more context: 1M tokens versus 131K.
  • Qwen2.5 7B Instruct has downloadable open weights; the other is API-only.

Side by side

Claude Sonnet 4.6Qwen2.5 7B Instruct
ProviderAnthropicAlibaba (Qwen)
Noometry Index50.128.6
Released2026-02-172024-09
WeightsProprietaryOpen
Context window1M131K
Max output128K8K
Input $ / M tokens$3$0.17
Output $ / M tokens$15$0.70
Results tracked5715

Sponsored placements are available on pages like this one. Advertise on Noometry

Category by category

Coding Claude Sonnet 4.6 leads

Claude Sonnet 4.6: 46.0 (#62), Qwen2.5 7B Instruct: 35.9 (#204)

BenchmarkClaude Sonnet 4.6Qwen2.5 7B Instruct
SWE-bench Verified75.2%—
DeepSWE29.9%—
FrontierCode24.3%—
LMArena WebDev1521—
SciCode46.8%—
WeirdML66.1%—
BigCodeBench Instruct—37.6%
LMArena Coding1504—
BigCodeBench Complete—46.1%
ALE-Bench1,327—

Agentic & Tool Use Claude Sonnet 4.6 leads

Claude Sonnet 4.6: 39.1 (#26), Qwen2.5 7B Instruct: 23.5 (#117)

BenchmarkClaude Sonnet 4.6Qwen2.5 7B Instruct
Terminal-Bench53.4%—
APEX-Agents43%—
OSWorld 2.09.3%—
DeepResearch Bench54.9%—
OSWorld72.1%—
BALROG—7.8%
ExploitBench23.6%—
GBAEval48.8%—
GDP.pdf18%—
LMArena Search1221—
Vending-Bench 27,204—

Reasoning Claude Sonnet 4.6 leads

Claude Sonnet 4.6: 46.1 (#40), Qwen2.5 7B Instruct: 14.3 (#317)

BenchmarkClaude Sonnet 4.6Qwen2.5 7B Instruct
Chess Puzzles13%0%
DTBench89.9%47.7%
LMCA46.5%6.4%
Epoch Capabilities Index152.25118.44
ARC-AGI-260.4%—
NYT Connections (extended)80.9%—
ARC-AGI-186.5%—
CritPt3.1%—
Thematic Generalization76.3%—
LMArena Hard Prompts1485—
Mystery Game Puzzles16%—
ForecastBench62—

Math Claude Sonnet 4.6 leads

Claude Sonnet 4.6: 51.9 (#41), Qwen2.5 7B Instruct: 11.8 (#300)

BenchmarkClaude Sonnet 4.6Qwen2.5 7B Instruct
OTIS Mock AIME 2024-202585.8%2.5%
ProofBench45%—
Omni-MATH—29.4%
LMArena Math1463—
FrontierMath (Feb 2025 set)32.4%—
FrontierMath Tier 4 (v1)8.3%—

Knowledge Claude Sonnet 4.6 leads

Claude Sonnet 4.6: 51.6 (#60), Qwen2.5 7B Instruct: 17.0 (#281)

BenchmarkClaude Sonnet 4.6Qwen2.5 7B Instruct
GPQA Diamond87.4%35.5%
SimpleQA Verified35.5%—
MMLU-Pro—53.9%
Vectara Hallucination Rate10.6%—
GPQA (HELM)—34.1%
LMArena Expert1500—
MMLU—72.9%

Multimodal Not comparable

Claude Sonnet 4.6: 38.0 (#61), Qwen2.5 7B Instruct: —

BenchmarkClaude Sonnet 4.6Qwen2.5 7B Instruct
LMArena Vision1283—
Blueprint-Bench 26.7%—
LMArena Document1482—

Multilingual Not comparable

Claude Sonnet 4.6: 54.6 (#39), Qwen2.5 7B Instruct: —

BenchmarkClaude Sonnet 4.6Qwen2.5 7B Instruct
LMArena Non-English1440—
LMArena Chinese1491—
LMArena French1465—
LMArena German1428—
LMArena Japanese1419—
LMArena Korean1411—
LMArena Russian1440—
LMArena Spanish1465—

Instruction Following Claude Sonnet 4.6 leads

Claude Sonnet 4.6: 77.6 (#23), Qwen2.5 7B Instruct: 63.4 (#226)

BenchmarkClaude Sonnet 4.6Qwen2.5 7B Instruct
IFEval—74.1%
LMArena Instruction Following1476—

Long Context Not comparable

Claude Sonnet 4.6: 45.3 (#40), Qwen2.5 7B Instruct: —

BenchmarkClaude Sonnet 4.6Qwen2.5 7B Instruct
LMArena Longer Query1479—

Writing & Preference Claude Sonnet 4.6 leads

Claude Sonnet 4.6: 70.4 (#22), Qwen2.5 7B Instruct: 48.8 (#194)

BenchmarkClaude Sonnet 4.6Qwen2.5 7B Instruct
LMArena Text1458—
LMArena Creative Writing1435—
EQ-Bench Creative Writing1810—
WildBench—73.1%
EQ-Bench 41207—
LMArena Multi-Turn1464—

Frequently asked questions

Is Claude Sonnet 4.6 better than Qwen2.5 7B Instruct?

Claude Sonnet 4.6 is the stronger model overall, scoring 50.1 to 28.6 on the Noometry Index. Qwen2.5 7B Instruct costs 20× less per token, which makes it the better buy when Claude Sonnet 4.6's lead doesn't matter for your workload.

Which is cheaper, Claude Sonnet 4.6 or Qwen2.5 7B Instruct?

Qwen2.5 7B Instruct is cheaper. It lists at $0.17 per million input tokens and $0.70 per million output tokens; Claude Sonnet 4.6 lists at $3 and $15.

Is Claude Sonnet 4.6 or Qwen2.5 7B Instruct better for coding?

Claude Sonnet 4.6 scores higher on coding benchmarks: 46.0 versus 35.9 in the Noometry coding category.

Which has the bigger context window?

Claude Sonnet 4.6 does, with 1M tokens against 131K.

How many benchmarks do Claude Sonnet 4.6 and Qwen2.5 7B Instruct share?

6 benchmarks have published results for both models. Claude Sonnet 4.6 has 57 scored results on Noometry and Qwen2.5 7B Instruct has 15.

Related comparisons

Go deeper