Noometry

Model comparison

Claude Opus 4.8 vs DeepSeek-V3

Claude Opus 4.8 is the stronger model overall, scoring 60.7 to 39.5 on the Noometry Index. DeepSeek-V3 costs 25× less per token, which makes it the better buy when Claude Opus 4.8's lead doesn't matter for your workload.

Last verified . 30 shared benchmarks.

Claude Opus 4.8 Anthropic

60.7

Rank #13 Confirmed

DeepSeek-V3 DeepSeek

39.5

Rank #166 Confirmed

Summary

  • They share 30 benchmarks with published results for both. Claude Opus 4.8 scores higher in 8 categories and DeepSeek-V3 in 0 categories; 8 gaps are clear of the uncertainty.
  • The widest gap is in math, where Claude Opus 4.8 leads 78.4 to 32.1.
  • The biggest single-benchmark swing is OTIS Mock AIME 2024-2025: 98.3% for Claude Opus 4.8 and 37.8% for DeepSeek-V3.
  • DeepSeek-V3 is cheaper at $0.24 / $0.90 per million input/output tokens, against $5 / $25 for Claude Opus 4.8.
  • Claude Opus 4.8 accepts more context: 1M tokens versus 164K.
  • DeepSeek-V3 has downloadable open weights; the other is API-only.

Side by side

Claude Opus 4.8DeepSeek-V3
ProviderAnthropicDeepSeek
Noometry Index60.739.5
Released2026-05-282024-12-26
WeightsProprietaryOpen
Context window1M164K
Max output128K164K
Input $ / M tokens$5$0.24
Output $ / M tokens$25$0.90
Results tracked6560

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: 59.9 (#12), DeepSeek-V3: 42.3 (#106)

BenchmarkClaude Opus 4.8DeepSeek-V3
SciCode53.5%35.8%
WeirdML82.9%36.1%
LMArena Coding14901368
DeepSWE59%—
FrontierCode46.5%—
Aider Polyglot—55.1%
LMArena WebDev1556—
GSO47.1%—
BigCodeBench Instruct—50%
LiveBench Coding—70.9%
BigCodeBench Complete—62.2%
ALE-Bench1,564—
HumanEval+—86.6%
MBPP+—73%

Agentic & Tool Use Not comparable

Claude Opus 4.8: 47.6 (#11), DeepSeek-V3: —

BenchmarkClaude Opus 4.8DeepSeek-V3
APEX-Agents48.9%—
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—49.6%
Vending-Bench 25,787—

Reasoning Claude Opus 4.8 leads

Claude Opus 4.8: 64.7 (#16), DeepSeek-V3: 20.5 (#236)

BenchmarkClaude Opus 4.8DeepSeek-V3
SimpleBench64.8%27.2%
Kagi LLM Benchmark88.8%52.3%
CritPt20.9%0%
LMArena Hard Prompts14821365
DTBench94.9%64.8%
LMCA57.5%15.5%
Epoch Capabilities Index158.21135.94
ForecastBench59.959.1
ARC-AGI-272.1%—
NYT Connections (extended)91.1%—
ARC-AGI-192.5%—
Chess Puzzles34%—
EnigmaEval23.5%—
EBR-Bench28.6%—
LiveBench Reasoning—65.8%
Mystery Game Puzzles36%—
LiveBench Data Analysis—60.9%
Surface Evolver Bench87.5%—
Bench to the Future 30.14—
BIG-Bench Hard—87.5%
HellaSwag—88.9%
LiveBench—66.9%
PIQA—84.7%
WinoGrande—85.2%

Math Claude Opus 4.8 leads

Claude Opus 4.8: 78.4 (#13), DeepSeek-V3: 32.1 (#219)

BenchmarkClaude Opus 4.8DeepSeek-V3
OTIS Mock AIME 2024-202598.3%37.8%
LMArena Math14871373
FrontierMath (Feb 2025 set)47.2%1.7%
FrontierMath (Tiers 1-3)80%—
FrontierMath Tier 456.1%—
MathArena Final-Answer Competitions91.8%—
ProofBench69%—
Omni-MATH—40.3%
LiveBench Math—73.5%
MATH Level 5—75.5%
FrontierMath Tier 4 (v1)31.3%—

Knowledge Claude Opus 4.8 leads

Claude Opus 4.8: 61.3 (#29), DeepSeek-V3: 37.5 (#155)

BenchmarkClaude Opus 4.8DeepSeek-V3
GPQA Diamond91%67.6%
LMArena Expert15021351
SimpleQA Verified53%—
MMLU-Pro—72.3%
Confabulations—26.1%
Vectara Hallucination Rate—6.1%
GPQA (HELM)—53.8%
ARC (AI2) Challenge—95.3%
MMLU—87.2%
TriviaQA—82.9%

Multimodal Not comparable

Claude Opus 4.8: 42.9 (#26), DeepSeek-V3: —

BenchmarkClaude Opus 4.8DeepSeek-V3
LMArena Vision1294—
Blueprint-Bench 214.5%—
Furniture Assembly42.5%—
LMArena Document1475—

Multilingual Claude Opus 4.8 leads

Claude Opus 4.8: 55.2 (#33), DeepSeek-V3: 48.5 (#143)

BenchmarkClaude Opus 4.8DeepSeek-V3
LMArena Non-English14501358
LMArena Chinese15071391
LMArena French14811385
LMArena German14721374
LMArena Japanese14401333
LMArena Korean14321319
LMArena Russian14741373
LMArena Spanish14661358

Instruction Following Claude Opus 4.8 leads

Claude Opus 4.8: 77.4 (#24), DeepSeek-V3: 72.8 (#130)

BenchmarkClaude Opus 4.8DeepSeek-V3
LMArena Instruction Following14761345
LiveBench Instruction Following—81.5%
IFEval—83.2%

Long Context Claude Opus 4.8 leads

Claude Opus 4.8: 45.4 (#35), DeepSeek-V3: 34.0 (#253)

BenchmarkClaude Opus 4.8DeepSeek-V3
LMArena Longer Query14831352
Fiction.LiveBench—50%

Writing & Preference Claude Opus 4.8 leads

Claude Opus 4.8: 72.0 (#16), DeepSeek-V3: 57.4 (#130)

BenchmarkClaude Opus 4.8DeepSeek-V3
LMArena Text14611375
LMArena Creative Writing14541364
EQ-Bench Creative Writing18401472
LMArena Multi-Turn14761389
Short-Story Creative Writing—77%
WildBench—83%
EQ-Bench 41281—
LiveBench Language—49.1%

Frequently asked questions

Is Claude Opus 4.8 better than DeepSeek-V3?

Claude Opus 4.8 is the stronger model overall, scoring 60.7 to 39.5 on the Noometry Index. DeepSeek-V3 costs 25× 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 DeepSeek-V3?

DeepSeek-V3 is cheaper. It lists at $0.24 per million input tokens and $0.90 per million output tokens; Claude Opus 4.8 lists at $5 and $25.

Is Claude Opus 4.8 or DeepSeek-V3 better for coding?

Claude Opus 4.8 scores higher on coding benchmarks: 59.9 versus 42.3 in the Noometry coding category.

Which has the bigger context window?

Claude Opus 4.8 does, with 1M tokens against 164K.

How many benchmarks do Claude Opus 4.8 and DeepSeek-V3 share?

30 benchmarks have published results for both models. Claude Opus 4.8 has 65 scored results on Noometry and DeepSeek-V3 has 60.

Related comparisons

Go deeper