ModelsCompareBest forBenchmarksStatusPricingAPI
Compare / head-to-head

Claude Haiku 4.5vsClaude Opus 4.6

Claude Opus 4.6 leads 4 of 4 shared benchmarks. Claude Haiku 4.5 is 5x cheaper per token. 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
0 – 4
Claude Opus 4.6 leads
Cheaper per token
Claude Haiku 4.5
5x cheaper, input + output
Larger context
Claude Opus 4.6
1M tokens
Provider uptime (30d)
100%
Anthropic
Claude Haiku 4.5
Anthropic · released 2025-10-15
textvision
Context
200K
Max out
64K
Input /1M
$1
Output /1M
$5
Cached /1M
$0.1
Scores
4 · 4 core
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
Quality

Benchmark matrix

BenchmarkClaude Haiku 4.5Claude Opus 4.6ΔEdge
Reported by both · 4
SWE-bench Verified73.3% 80.8% -7.5 ptClaude Opus 4.6
GPQA Diamond71.2% epoch91.3% -20.1 ptClaude Opus 4.6
OTIS Mock AIME 2024-2025 (Epoch AI run)66.7% epoch94.4% epoch-27.7 ptClaude Opus 4.6
SimpleQA Verified13.2% epoch47% epoch-33.8 ptClaude Opus 4.6
Only Claude Haiku 4.5 reports · 0
Claude Haiku 4.5 reports nothing Claude Opus 4.6 does not.
Only Claude Opus 4.6 reports · 14
AIME 2026not reported96.7% matharena
LMArena Elonot reported1497.5
ARC-AGI-2 (Verified)not reported68.8%
FrontierMath Tier 4 v2 (Epoch AI run)not reported26.8% epoch
FrontierMath Tiers 1-3 v2 (Epoch AI run)not reported66% epoch
MCP Atlasnot reported59.5%
MMMLUnot reported91.1%
MMMU-Pro (no tools)not reported73.9%
MMMU-Pro (with tools)not reported77.3%
OSWorld-Verifiednot reported72.7%
SWE-bench Verified (Epoch AI run)not reported78.7% epoch
Tau2-bench Retailnot reported91.9%
Tau2-bench Telecomnot reported99.3%
Terminal-Bench 2.0not reported65.4%
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 Haiku 4.5 minus Claude Opus 4.6 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 Haiku 4.5Claude Opus 4.6Edge
Context window200K tokens1M tokensClaude Opus 4.6
Max output64K tokens128K tokensClaude Opus 4.6
Input price / 1M$1 $5 Claude Haiku 4.5
Output price / 1M$5 $25 Claude Haiku 4.5
Cached input / 1M$0.1 $0.5 Claude Haiku 4.5
Input + output / 1M
Lower is cheaper. List prices; batch, tool and regional fees excluded.
$6$30Claude Haiku 4.5
Modalitiestext · visiontext · visionTie
Released2025-10-152026-02-05
Cited benchmark scores419
Reliability

Anthropic status

All providers
More matchups

Claude Haiku 4.5 vs …

More matchups

Claude Opus 4.6 vs …

Built by Respan
Which one wins on your data?

Public benchmarks are a starting point. Run Claude Haiku 4.5 and Claude Opus 4.6 on your own prompts with Respan evals, or route to either through one gateway key with automatic failover.