Compare / head-to-head
Claude Opus 4.5vs
GPT-5.6 Sol
GPT-5.6 Sol leads 24 of 24 shared benchmarks. GPT-5.6 Sol is 1.3x cheaper per token. GPT-5.6 Sol has the larger context window (1.1M tokens).
Benchmarks from cited public sources; pricing from official pages; status from official provider feeds.
Shared benchmarks
0 – 24
GPT-5.6 Sol leads
Cheaper per token
GPT-5.6 Sol
1.3x cheaper, input + output
Larger context
GPT-5.6 Sol
1.1M tokens
Providers
2 providers
Anthropic · OpenAI
Claude Opus 4.5
Anthropic · released 2025-11-24
textvision
GPT-5.6 Sol
OpenAI · released 2026-07-09
textvision
Quality
Benchmark matrix
24 shared · 14 only Claude Opus 4.5 · 10 only GPT-5.6 Sol| Benchmark | Claude Opus 4.5 | GPT-5.6 Sol | Δ | Edge |
|---|---|---|---|---|
| Reported by both · 24 | ||||
| GPQA Diamond | 87% ↗ | 94.6% ↗ | -7.6 pt | GPT-5.6 Sol |
| LMArena Elo | 1473.6 ↗ | 1483.9 ↗ | -10.3 | GPT-5.6 Sol |
| BALROG (BALROG) | 43.5% ↗ | 60% ↗ | -16.5 pt | GPT-5.6 Sol |
| Chess Puzzles (Epoch AI run) | 12% ↗epoch run | 64% ↗epoch run | -52 pt | GPT-5.6 Sol |
| EBR-bench (Epoch AI run) | 14.3% ↗epoch run | 44.8% ↗epoch run | -30.5 pt | GPT-5.6 Sol |
| FrontierMath Tier 4 v2 (Epoch AI run) | 4.9% ↗epoch run | 82.9% ↗epoch run | -78 pt | GPT-5.6 Sol |
| FrontierMath Tiers 1-3 v2 (Epoch AI run) | 34.4% ↗epoch run | 89.1% ↗epoch run | -54.7 pt | GPT-5.6 Sol |
| Furniture Assembly (Epoch AI run) | 28.3% ↗epoch run | 56.7% ↗epoch run | -28.4 pt | GPT-5.6 Sol |
| GSO Opt@1 (GSO) | 24.5% ↗ | 70.6% ↗ | -46.1 pt | GPT-5.6 Sol |
| LiveBench Agentic Coding (LiveBench) | 39.7% ↗ | 56.2% ↗ | -16.5 pt | GPT-5.6 Sol |
| LiveBench Coding (LiveBench) | 79.7% ↗ | 83.9% ↗ | -4.2 pt | GPT-5.6 Sol |
| LiveBench Data Analysis (LiveBench) | 74.4% ↗ | 79.8% ↗ | -5.4 pt | GPT-5.6 Sol |
| LiveBench Instruction Following (LiveBench) | 62.5% ↗ | 71.8% ↗ | -9.3 pt | GPT-5.6 Sol |
| LiveBench Language (LiveBench) | 81.3% ↗ | 87.7% ↗ | -6.4 pt | GPT-5.6 Sol |
| LiveBench Mathematics (LiveBench) | 90.4% ↗ | 96.2% ↗ | -5.8 pt | GPT-5.6 Sol |
| LiveBench Reasoning (LiveBench) | 80.1% ↗ | 91.7% ↗ | -11.6 pt | GPT-5.6 Sol |
| Mystery Game Puzzles (Epoch AI run) | 22% ↗epoch run | 58% ↗epoch run | -36 pt | GPT-5.6 Sol |
| OTIS Mock AIME 2024-2025 (Epoch AI run) | 86.1% ↗epoch run | 100% ↗epoch run | -13.9 pt | GPT-5.6 Sol |
| SAGE (Vals AI) | 52.1% ↗ | 52.6% ↗ | -0.5 pt | GPT-5.6 Sol |
| SimpleBench (SimpleBench) | 62% ↗ | 71.7% ↗ | -9.7 pt | GPT-5.6 Sol |
| SimpleQA Verified | 45.7% ↗epoch run | 69.7% ↗epoch run | -24 pt | GPT-5.6 Sol |
| tau2-bench Banking Knowledge (Sierra) | 24.7% ↗ | 46.9% ↗ | -22.2 pt | GPT-5.6 Sol |
| Vending-Bench 2 (Andon Labs) | 4967.06 ↗ | 9619.37 ↗ | -4652.3 | GPT-5.6 Sol |
| WeirdML (Håvard Tveit Ihle) | 63.7% ↗ | 89.4% ↗ | -25.7 pt | GPT-5.6 Sol |
| Only Claude Opus 4.5 reports · 14 | ||||
| SWE-bench Verified | 80.9% ↗ | not reported | — | — |
| ARC-AGI-2 (Verified) | 37.6% ↗ | not reported | — | — |
| LMArena WebDev (LMArena) | 1493.3 ↗ | not reported | — | — |
| MCP Atlas | 62.3% ↗ | not reported | — | — |
| MMMLU | 90.8% ↗ | not reported | — | — |
| MMMU (validation) | 80.7% ↗ | not reported | — | — |
| OSWorld | 66.3% ↗ | not reported | — | — |
| SWE-bench Verified (Epoch AI run) | 76.7% ↗epoch run | not reported | — | — |
| tau2-bench Airline (Sierra) | 84% ↗ | not reported | — | — |
| tau2-bench Retail (Sierra) | 79.6% ↗ | not reported | — | — |
| tau2-bench Telecom (Sierra) | 92.3% ↗ | not reported | — | — |
| Terminal-Bench 2.0 | 59.3% ↗ | not reported | — | — |
| τ2-bench (Retail) | 88.9% ↗ | not reported | — | — |
| τ2-bench (Telecom) | 98.2% ↗ | not reported | — | — |
| Only GPT-5.6 Sol reports · 10 | ||||
| APEX-Agents (Mercor) | not reported | 51.4% ↗ | — | — |
| DeepSWE v1.1 | not reported | 72.7% ↗ | — | — |
| DeepSWE v1.1 (Datacurve) | not reported | 72.7% ↗ | — | — |
| FrontierSWE V2 (Proximal Labs) | not reported | 32.2% ↗ | — | — |
| LMArena Agent (LMArena) | not reported | 0.0645 ↗ | — | — |
| LMArena Vision (LMArena) | not reported | 1284.7 ↗ | — | — |
| MirrorCode (Epoch AI run) | not reported | 20% ↗epoch run | — | — |
| SWE-Bench Pro | not reported | 64.6% ↗ | — | — |
| SWE-rebench 2026-05-15 to 2026-07-01 (Nebius) | not reported | 62.3% ↗ | — | — |
| Terminal-Bench 4.0 (Vals AI) | not reported | 37.9% ↗ | — | — |
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 4.5 minus GPT-5.6 Sol 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 4.5 | GPT-5.6 Sol | Edge |
|---|---|---|---|
| Context window | 200K tokens | 1.1M tokens | GPT-5.6 Sol |
| Max output | 64K tokens | 128K tokens | GPT-5.6 Sol |
| Input price / 1M | $5 ↗ | $4 ↗ | GPT-5.6 Sol |
| Output price / 1M | $25 ↗ | $20 ↗ | GPT-5.6 Sol |
| Cached input / 1M | $0.5 ↗ | $0.4 ↗ | GPT-5.6 Sol |
| Input + output / 1M Lower is cheaper. List prices; batch, tool and regional fees excluded. | $30 | $24 | GPT-5.6 Sol |
| Modalities | text · vision | text · vision | Tie |
| Released | 2025-11-24 | 2026-07-09 | — |
| Cited benchmark scores | 39 | 40 | — |
Reliability
Provider status
Live from /statusMore matchups
Claude Opus 4.5 vs …
Models sharing the most benchmarksMore matchups
GPT-5.6 Sol vs …
Models sharing the most benchmarksBuilt by Respan
Which one wins on your data?
Public benchmarks are a starting point. Run Claude Opus 4.5 and GPT-5.6 Sol on your own prompts with Respan evals, or route to either through one gateway key with automatic failover.