ModelsCompareBest forBenchmarksStatusPricingAPI
Compare / head-to-head

Claude Opus 4.6vsHy3

Claude Opus 4.6 leads 3 of 4 shared benchmarks. Claude Opus 4.6 has the larger context window (1M tokens).

Benchmarks from cited public sources; pricing from official pages; status from official provider feeds.

Shared benchmarks
3 – 1
Claude Opus 4.6 leads
Cheaper per token
list price, input + output
Larger context
Claude Opus 4.6
1M tokens
Provider uptime (30d)
100% · —
Anthropic · Tencent Hunyuan
Claude Opus 4.6
Anthropic · released 2026-02-05
textvision
Context
1M
Max out
128K
Input /1M
$5
Output /1M
$25
Cached /1M
$0.5
Scores
19 · 9 core
Hy3
Tencent Hunyuan · released 2026-07-06
text
Context
256K
Max out
Input /1M
Output /1M
Cached /1M
Scores
41 · 8 core
Quality

Benchmark matrix

BenchmarkClaude Opus 4.6Hy3ΔEdge
Reported by both · 4
SWE-bench Verified80.8% 78% +2.8 ptClaude Opus 4.6
GPQA Diamond91.3% 90.4% +0.9 ptClaude Opus 4.6
LMArena Elo1497.5 1440.6 +56.9Claude Opus 4.6
MCP Atlas59.5% 79.1% -19.6 ptHy3
Only Claude Opus 4.6 reports · 14
AIME 202696.7% matharenanot reported
ARC-AGI-2 (Verified)68.8% not reported
FrontierMath Tier 4 v2 (Epoch AI run)26.8% epochnot reported
FrontierMath Tiers 1-3 v2 (Epoch AI run)66% epochnot reported
MMMLU91.1% not reported
MMMU-Pro (no tools)73.9% not reported
MMMU-Pro (with tools)77.3% not reported
OSWorld-Verified72.7% not reported
OTIS Mock AIME 2024-2025 (Epoch AI run)94.4% epochnot reported
SimpleQA Verified47% epochnot reported
SWE-bench Verified (Epoch AI run)78.7% epochnot reported
Tau2-bench Retail91.9% not reported
Tau2-bench Telecom99.3% not reported
Terminal-Bench 2.065.4% not reported
Only Hy3 reports · 37
Humanity's Last Exam (no tools)not reported37%
AA-LCRnot reported73.4%
Apex-Agentnot reported25.6%
ArxivMathnot reported52.2%
BrowseCompnot reported84.2%
CL-benchnot reported23.8%
CL-bench Lifenot reported17%
ClawEvalnot reported68.5%
CMT-Benchmarknot reported37.8%
DeepSearchQAnot reported91%
DeepSWEnot reported28%
e-benchnot reported50.2%
FrontierScience-Olympiadnot reported74.8%
FrontierScience-Researchnot reported21.3%
HLE (with tools, text-only)not reported53.2%
HorizonMathnot reported7.1%
Hy-Backend 2.0not reported25%
Hy-CompanyBenchnot reported41.7%
Hy-Euler Pronot reported24.2%
Hy-FinModelBenchnot reported69%
Hy-Mathnot reported60.9%
Hy-SkillsWorldnot reported45.8%
Hy-SWE Maxnot reported49%
IMOAnswerBenchnot reported90%
MathArena Apexnot reported38.7%
NL2Reponot reported45.6%
PHYBenchnot reported77.4%
ProdBenchnot reported23%
SkillsBenchnot reported55.3%
SuperChemnot reported54.9%
SWE-bench Multilingualnot reported75.8%
SWE-bench Pronot reported57.9%
Terminal-Bench 2.1not reported71.7%
Toolathlonnot reported48.5%
USAMO 2026not reported72%
WideSearchnot reported76.4%
WildClawBenchnot reported53.6%
Scores tagged "epoch" or "matharena" are independent runs, used only where the lab has not published its own; ⚠ marks rows MathArena flags as released after the competition. Higher is better on every row. Δ is Claude Opus 4.6 minus Hy3 in the benchmark's own unit. "Not reported" means the lab has not published that figure; it is not a zero. ↗ opens the source.
Specs & pricing

Side by side

SpecClaude Opus 4.6Hy3Edge
Context window1M tokens256K tokensClaude Opus 4.6
Max output128K tokens
Input price / 1M$5
Output price / 1M$25
Cached input / 1M$0.5
Input + output / 1M
Lower is cheaper. List prices; batch, tool and regional fees excluded.
$30
Modalitiestext · visiontextClaude Opus 4.6
Released2026-02-052026-07-06
Cited benchmark scores1941
Reliability

Provider status

All providers
More matchups

Claude Opus 4.6 vs …

Built by Respan
Which one wins on your data?

Public benchmarks are a starting point. Run Claude Opus 4.6 and Hy3 on your own prompts with Respan evals, or route to either through one gateway key with automatic failover.