Noometry

Model comparison

Claude Opus 4.5 vs gpt-oss-20b

Claude Opus 4.5 is the stronger model overall, scoring 50.5 to 32.5 on the Noometry Index. gpt-oss-20b costs 278× less per token, which makes it the better buy when Claude Opus 4.5's lead doesn't matter for your workload.

Last verified . 27 shared benchmarks.

Claude Opus 4.5 Anthropic

50.5

Rank #47 Confirmed

gpt-oss-20b OpenAI

32.5

Rank #255 Confirmed

Summary

  • They share 27 benchmarks with published results for both. Claude Opus 4.5 scores higher in 8 categories and gpt-oss-20b in 1 category; 8 gaps are clear of the uncertainty.
  • The widest gap is in agentic & tool use, where Claude Opus 4.5 leads 47.3 to 9.3.
  • The biggest single-benchmark swing is Terminal-Bench: 63.1% for Claude Opus 4.5 and 3.4% for gpt-oss-20b.
  • gpt-oss-20b is cheaper at $0.018 / $0.09 per million input/output tokens, against $5 / $25 for Claude Opus 4.5.
  • Claude Opus 4.5 accepts more context: 200K tokens versus 131K.
  • gpt-oss-20b has downloadable open weights; the other is API-only.

Side by side

Claude Opus 4.5gpt-oss-20b
ProviderAnthropicOpenAI
Noometry Index50.532.5
Released2025-11-012025-08-05
WeightsProprietaryOpen
Context window200K131K
Max output64K16K
Input $ / M tokens$5$0.018
Output $ / M tokens$25$0.09
Results tracked6934

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

Category by category

Coding Claude Opus 4.5 leads

Claude Opus 4.5: 54.8 (#27), gpt-oss-20b: 37.6 (#192)

BenchmarkClaude Opus 4.5gpt-oss-20b
WeirdML63.7%40.9%
LMArena Coding15041306
ALE-Bench1,025566.05
SWE-bench Verified76.7%—
SWE-bench Verified (bash only)76.8%—
LMArena WebDev1494—
SWE-bench Multilingual70.7%—
SciCode—34.4%
GSO26.5%—
AlgoTune1.77—

Agentic & Tool Use Claude Opus 4.5 leads

Claude Opus 4.5: 47.3 (#12), gpt-oss-20b: 9.3 (#154)

BenchmarkClaude Opus 4.5gpt-oss-20b
Terminal-Bench63.1%3.4%
Berkeley Function Calling Leaderboard77.5%—
GDPval45.5%—
Remote Labor Index3.8%—
τ²-bench Airline84%—
τ²-bench Banking24.7%—
τ²-bench Retail79.6%—
τ²-bench Telecom92.3%—
Cybench82%—
DeepResearch Bench54.8%—
OSWorld66.3%—
BALROG43.5%—
LMArena Search1180—
METR Time Horizons75%—
Vending-Bench 24,967—

Reasoning Claude Opus 4.5 leads

Claude Opus 4.5: 42.6 (#51), gpt-oss-20b: 19.3 (#261)

BenchmarkClaude Opus 4.5gpt-oss-20b
Kagi LLM Benchmark80.2%53.2%
Chess Puzzles12%4%
LMArena Hard Prompts14761274
DTBench89.9%68%
LMCA44.5%14.5%
Epoch Capabilities Index150.09137.82
ARC-AGI-237.6%—
SimpleBench62%—
NYT Connections (extended)52.5%—
ARC-AGI-180%—
CritPt—1.4%
EnigmaEval11.9%—
EBR-Bench14.3%—
Mystery Game Puzzles22%—
ForecastBench60.7—

Math Too close to call

Claude Opus 4.5: 38.6 (#132), gpt-oss-20b: 39.4 (#103)

BenchmarkClaude Opus 4.5gpt-oss-20b
OTIS Mock AIME 2024-202586.1%65.3%
LMArena Math14631317
FrontierMath (Tiers 1-3)34.4%—
FrontierMath Tier 44.9%—
ProofBench36%—
Omni-MATH—56.5%
FrontierMath (Feb 2025 set)20.7%—
FrontierMath Tier 4 (v1)4.2%—

Knowledge Claude Opus 4.5 leads

Claude Opus 4.5: 56.5 (#44), gpt-oss-20b: 34.6 (#195)

BenchmarkClaude Opus 4.5gpt-oss-20b
GPQA Diamond86%60.8%
LMArena Expert14871258
Humanity's Last Exam25.2%—
SimpleQA Verified45.7%—
MMLU-Pro—74%
Vectara Hallucination Rate10.9%—
GPQA (HELM)—59.4%

Multimodal Not comparable

Claude Opus 4.5: 31.4 (#107), gpt-oss-20b: —

BenchmarkClaude Opus 4.5gpt-oss-20b
GeoBench75%—
VPCT40%—
Furniture Assembly28.3%—
LMArena Document1462—

Multilingual Claude Opus 4.5 leads

Claude Opus 4.5: 54.3 (#47), gpt-oss-20b: 42.2 (#197)

BenchmarkClaude Opus 4.5gpt-oss-20b
LMArena Non-English14381268
LMArena Chinese14701314
LMArena German14491255
LMArena Japanese14161244
LMArena Korean14241236
LMArena Russian14471278
LMArena Spanish14581267
LMArena French1471—

Instruction Following Claude Opus 4.5 leads

Claude Opus 4.5: 77.5 (#19), gpt-oss-20b: 61.8 (#240)

BenchmarkClaude Opus 4.5gpt-oss-20b
LMArena Instruction Following14781236
IFEval—73.2%

Long Context Claude Opus 4.5 leads

Claude Opus 4.5: 46.5 (#22), gpt-oss-20b: 37.9 (#209)

BenchmarkClaude Opus 4.5gpt-oss-20b
LMArena Longer Query14801250
CL-bench21.1%—

Writing & Preference Claude Opus 4.5 leads

Claude Opus 4.5: 68.1 (#28), gpt-oss-20b: 35.5 (#265)

BenchmarkClaude Opus 4.5gpt-oss-20b
LMArena Text14511287
LMArena Creative Writing14451201
EQ-Bench Creative Writing1687666
LMArena Multi-Turn14661268
WildBench—73.7%

Frequently asked questions

Is Claude Opus 4.5 better than gpt-oss-20b?

Claude Opus 4.5 is the stronger model overall, scoring 50.5 to 32.5 on the Noometry Index. gpt-oss-20b costs 278× less per token, which makes it the better buy when Claude Opus 4.5's lead doesn't matter for your workload.

Which is cheaper, Claude Opus 4.5 or gpt-oss-20b?

gpt-oss-20b is cheaper. It lists at $0.018 per million input tokens and $0.09 per million output tokens; Claude Opus 4.5 lists at $5 and $25.

Is Claude Opus 4.5 or gpt-oss-20b better for coding?

Claude Opus 4.5 scores higher on coding benchmarks: 54.8 versus 37.6 in the Noometry coding category.

Which has the bigger context window?

Claude Opus 4.5 does, with 200K tokens against 131K.

How many benchmarks do Claude Opus 4.5 and gpt-oss-20b share?

27 benchmarks have published results for both models. Claude Opus 4.5 has 69 scored results on Noometry and gpt-oss-20b has 34.

Related comparisons

Go deeper