Noometry

Model comparison

Claude Opus 4.8 vs gpt-oss-120b

Claude Opus 4.8 is the stronger model overall, scoring 60.7 to 36.0 on the Noometry Index. gpt-oss-120b costs 142× less per token, which makes it the better buy when Claude Opus 4.8's lead doesn't matter for your workload.

Last verified . 34 shared benchmarks.

Claude Opus 4.8 Anthropic

60.7

Rank #9 Confirmed

gpt-oss-120b OpenAI

36.0

Rank #210 Confirmed

Summary

  • They share 34 benchmarks with published results for both. Claude Opus 4.8 scores higher in 9 categories and gpt-oss-120b in 0 categories; 9 gaps are clear of the uncertainty.
  • The widest gap is in reasoning, where Claude Opus 4.8 leads 64.6 to 19.5.
  • The biggest single-benchmark swing is Surface Evolver Bench: 87.5% for Claude Opus 4.8 and 25% for gpt-oss-120b.
  • gpt-oss-120b is cheaper at $0.037 / $0.17 per million input/output tokens, against $5 / $25 for Claude Opus 4.8.
  • Claude Opus 4.8 accepts more context: 1M tokens versus 131K.
  • gpt-oss-120b has downloadable open weights; the other is API-only.

Side by side

Claude Opus 4.8gpt-oss-120b
ProviderAnthropicOpenAI
Noometry Index60.736.0
Released2026-05-282025-08-05
WeightsProprietaryOpen
Context window1M131K
Max output128K41K
Input $ / M tokens$5$0.037
Output $ / M tokens$25$0.17
Results tracked6548

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

Category by category

Coding Claude Opus 4.8 leads

Claude Opus 4.8: 60.2 (#9), gpt-oss-120b: 33.2 (#251)

BenchmarkClaude Opus 4.8gpt-oss-120b
SciCode53.5%36%
WeirdML82.9%48.2%
LMArena Coding14911380
ALE-Bench1,564575.62
DeepSWE59%—
FrontierCode46.5%—
SWE-bench Verified (bash only)—26%
Aider Polyglot—41.8%
LMArena WebDev1556—
GSO47.1%—
AlgoTune—1.41

Agentic & Tool Use Claude Opus 4.8 leads

Claude Opus 4.8: 47.6 (#10), gpt-oss-120b: 12.1 (#145)

BenchmarkClaude Opus 4.8gpt-oss-120b
APEX-Agents48.9%4.4%
Vending-Bench 25,787-21.53
Terminal-Bench—18.7%
OSWorld 2.020.6%—
Remote Labor Index8.3%—
τ²-bench Banking39.7%—
DeepResearch Bench50.2%—
PostTrainBench33.8%—
GBAEval70.9%—
GDP.pdf24%—
LMArena Search1204—
METR Time Horizons—56.6%

Reasoning Claude Opus 4.8 leads

Claude Opus 4.8: 64.6 (#14), gpt-oss-120b: 19.5 (#236)

BenchmarkClaude Opus 4.8gpt-oss-120b
SimpleBench64.8%22.1%
Kagi LLM Benchmark88.8%58.6%
CritPt20.9%1.1%
Chess Puzzles34%20%
LMArena Hard Prompts14811364
Mystery Game Puzzles36%2%
DTBench94.9%76.3%
LMCA57.5%22.1%
Surface Evolver Bench87.5%25%
Epoch Capabilities Index158.3139.94
ARC-AGI-272.1%—
NYT Connections (extended)91.1%—
ARC-AGI-192.5%—
EnigmaEval23.5%—
EBR-Bench28.6%—
Bench to the Future 30.14—
ForecastBench59.9—

Math Claude Opus 4.8 leads

Claude Opus 4.8: 78.2 (#11), gpt-oss-120b: 50.6 (#43)

BenchmarkClaude Opus 4.8gpt-oss-120b
OTIS Mock AIME 2024-202598.3%88.9%
LMArena Math14871388
FrontierMath (Tiers 1-3)80%—
FrontierMath Tier 456.1%—
MathArena Final-Answer Competitions91.8%—
ProofBench69%—
Omni-MATH—68.8%
FrontierMath (Feb 2025 set)47.2%—
FrontierMath Tier 4 (v1)31.3%—

Knowledge Claude Opus 4.8 leads

Claude Opus 4.8: 61.3 (#26), gpt-oss-120b: 42.3 (#89)

BenchmarkClaude Opus 4.8gpt-oss-120b
GPQA Diamond91%75.8%
LMArena Expert15011356
SimpleQA Verified53%—
MMLU-Pro—79.5%
Confabulations—15.7%
Vectara Hallucination Rate—14.2%
GPQA (HELM)—68.4%

Multimodal Not comparable

Claude Opus 4.8: 43.2 (#20), gpt-oss-120b: —

BenchmarkClaude Opus 4.8gpt-oss-120b
LMArena Vision1295—
Blueprint-Bench 214.5%—
Furniture Assembly42.5%—
LMArena Document1475—

Multilingual Claude Opus 4.8 leads

Claude Opus 4.8: 55.4 (#31), gpt-oss-120b: 48.2 (#142)

BenchmarkClaude Opus 4.8gpt-oss-120b
LMArena Non-English14501350
LMArena Chinese15091384
LMArena French14821370
LMArena German14701355
LMArena Japanese14401332
LMArena Korean14321283
LMArena Russian14761343
LMArena Spanish14701388

Instruction Following Claude Opus 4.8 leads

Claude Opus 4.8: 77.6 (#22), gpt-oss-120b: 69.5 (#168)

BenchmarkClaude Opus 4.8gpt-oss-120b
LMArena Instruction Following14761318
IFEval—83.6%

Long Context Claude Opus 4.8 leads

Claude Opus 4.8: 45.4 (#37), gpt-oss-120b: 31.4 (#273)

BenchmarkClaude Opus 4.8gpt-oss-120b
LMArena Longer Query14821319
Fiction.LiveBench—44.4%

Writing & Preference Claude Opus 4.8 leads

Claude Opus 4.8: 72.1 (#17), gpt-oss-120b: 46.7 (#213)

BenchmarkClaude Opus 4.8gpt-oss-120b
LMArena Text14601366
LMArena Creative Writing14531274
EQ-Bench Creative Writing1840961
LMArena Multi-Turn14761340
Short-Story Creative Writing—77.1%
WildBench—84.5%
EQ-Bench 41281—

Frequently asked questions

Is Claude Opus 4.8 better than gpt-oss-120b?

Claude Opus 4.8 is the stronger model overall, scoring 60.7 to 36.0 on the Noometry Index. gpt-oss-120b costs 142× less per token, which makes it the better buy when Claude Opus 4.8's lead doesn't matter for your workload.

Which is cheaper, Claude Opus 4.8 or gpt-oss-120b?

gpt-oss-120b is cheaper. It lists at $0.037 per million input tokens and $0.17 per million output tokens; Claude Opus 4.8 lists at $5 and $25.

Is Claude Opus 4.8 or gpt-oss-120b better for coding?

Claude Opus 4.8 scores higher on coding benchmarks: 60.2 versus 33.2 in the Noometry coding category.

Which has the bigger context window?

Claude Opus 4.8 does, with 1M tokens against 131K.

How many benchmarks do Claude Opus 4.8 and gpt-oss-120b share?

34 benchmarks have published results for both models. Claude Opus 4.8 has 65 scored results on Noometry and gpt-oss-120b has 48.

Related comparisons

Go deeper