Compare / head-to-head
Claude Opus 4.6vsStep 3.7 Flash
Claude Opus 4.6 leads 2 of 2 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
2 – 0
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 · StepFun
Claude Opus 4.6
Anthropic · released 2026-02-05
textvision
Step 3.7 Flash
StepFun · released 2026-05-29
textvision
- Context
- 256K
- Max out
- —
- Input /1M
- —
- Output /1M
- —
- Cached /1M
- —
- Scores
- 19 · 6 core
Quality
Benchmark matrix
2 shared · 16 only Claude Opus 4.6 · 17 only Step 3.7 Flash| Benchmark | Claude Opus 4.6 | Step 3.7 Flash | Δ | Edge |
|---|---|---|---|---|
| Reported by both · 2 | ||||
| SWE-bench Verified | 80.8% ↗ | 76.5% ↗ | +4.3 pt | Claude Opus 4.6 |
| AIME 2026 | 96.7% ↗matharena | 95% ↗matharena ⚠ | +1.7 pt | Claude Opus 4.6 |
| Only Claude Opus 4.6 reports · 16 | ||||
| GPQA Diamond | 91.3% ↗ | not reported | — | — |
| LMArena Elo | 1497.5 ↗ | not reported | — | — |
| ARC-AGI-2 (Verified) | 68.8% ↗ | not reported | — | — |
| FrontierMath Tier 4 v2 (Epoch AI run) | 26.8% ↗epoch | not reported | — | — |
| FrontierMath Tiers 1-3 v2 (Epoch AI run) | 66% ↗epoch | not reported | — | — |
| MCP Atlas | 59.5% ↗ | not reported | — | — |
| MMMLU | 91.1% ↗ | not reported | — | — |
| MMMU-Pro (no tools) | 73.9% ↗ | not reported | — | — |
| MMMU-Pro (with tools) | 77.3% ↗ | not reported | — | — |
| OSWorld-Verified | 72.7% ↗ | not reported | — | — |
| OTIS Mock AIME 2024-2025 (Epoch AI run) | 94.4% ↗epoch | not reported | — | — |
| SimpleQA Verified | 47% ↗epoch | not reported | — | — |
| SWE-bench Verified (Epoch AI run) | 78.7% ↗epoch | not reported | — | — |
| Tau2-bench Retail | 91.9% ↗ | not reported | — | — |
| Tau2-bench Telecom | 99.3% ↗ | not reported | — | — |
| Terminal-Bench 2.0 | 65.4% ↗ | not reported | — | — |
| Only Step 3.7 Flash reports · 17 | ||||
| AA-LCR | not reported | 63.9% ↗ | — | — |
| BrowseComp | not reported | 75.8% ↗ | — | — |
| ClawEval v1.1 | not reported | 67.1% ↗ | — | — |
| DeepSearchQA Accuracy | not reported | 81.7% ↗ | — | — |
| DeepSearchQA F1 | not reported | 92.8% ↗ | — | — |
| GDPval-Stirrup | not reported | 1415.8 ↗ | — | — |
| GDPval-Stirrup Intelligence Index | not reported | 45.8 ↗ | — | — |
| HLE (with tools, text-only) | not reported | 49.7% ↗ | — | — |
| HLE (with tools) | not reported | 47.2% ↗ | — | — |
| ResearchRubrics | not reported | 71.7% ↗ | — | — |
| SimpleVQA (with tools) | not reported | 79.2% ↗ | — | — |
| SWE-bench Multilingual | not reported | 72.4% ↗ | — | — |
| SWE-bench Pro | not reported | 56.3% ↗ | — | — |
| Terminal-Bench 2.1 | not reported | 59.6% ↗ | — | — |
| Toolathlon | not reported | 49.5% ↗ | — | — |
| V* (with Python) | not reported | 95.3% ↗ | — | — |
| WorldVQA (with visual search) | not reported | 58.1% ↗ | — | — |
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 Step 3.7 Flash 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
Official pricing pages and model cards| Spec | Claude Opus 4.6 | Step 3.7 Flash | Edge |
|---|---|---|---|
| Context window | 1M tokens | 256K tokens | Claude Opus 4.6 |
| Max output | 128K 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 | — | — |
| Modalities | text · vision | text · vision | Tie |
| Released | 2026-02-05 | 2026-05-29 | — |
| Cited benchmark scores | 19 | 19 | — |
Reliability
Provider status
Last 60 days · refreshed daily on this page · live on /statusMore matchups
Claude Opus 4.6 vs …
Models sharing the most benchmarksMore matchups
Step 3.7 Flash vs …
Models sharing the most benchmarksBuilt by Respan
Which one wins on your data?
Public benchmarks are a starting point. Run Claude Opus 4.6 and Step 3.7 Flash on your own prompts with Respan evals, or route to either through one gateway key with automatic failover.