Compare / GPT-5 vs GPT-5.6 Sol
GPT-5 vs GPT-5.6 Sol
Which is cheaper, which is more capable, and which does better at each kind of work, from 24 results on 5 sources that measured both models.
Of the 24 results both models have, the better value on each. Results whose reported ranges overlap are too close to call.
Results as of 19 November 2026 (catalogue release 2026-11-19-3608133e8331).
GPT-5 or GPT-5.6 Sol? The short answer
Pick GPT-5.6 Sol if quality matters most: it scores higher (Intelligence Index 47.0 against 23.0). Pick GPT-5 if cost matters more: it costs 2.3 times less per token.
- GPT-5 is cheaper: $3.44 against $8.00 per million tokens, 2.3 times less.
- GPT-5.6 Sol scores higher on the Artificial Analysis Intelligence Index: 47.0 against 23.0.
- GPT-5.6 Sol is newer: released 9 Jul 2026, against 7 Aug 2025 for GPT-5.
- On third-party benchmarks, on the results both models have: GPT-5.6 Sol does better for agents and tool use, reasoning and knowledge, writing and sticking to the facts.
Head to head
Every fact side by side
| GPT-5 | GPT-5.6 Sol | |
|---|---|---|
| Price per million tokens | $1.25 in · $10.00 out | $4.00 in · $20.00 out |
| Blended (3 in : 1 out) | $3.44 | $8.00 |
| Cheapest host, blended | $3.44 (Azure) | $8.00 (Amazon Bedrock) |
| Cached input, per million | – | $0.40 cached input |
| Intelligence Index | 23.0 | 47.0 |
| Spring Prompt overall | – | – |
| Output speed, tokens a second | – | – |
| Context window | 400,000 tokens | 1,050,000 tokens |
| Released | 7 Aug 2025 | 9 Jul 2026 (newer) |
| Weights | Closed | Closed |
| Developer based in | the US | the US |
Prices are the developer's list price where we have read it, otherwise the typical host on OpenRouter; see every model's price. Intelligence and speed from Artificial Analysis.
Which is better for what
Each model's rank among the models on sale today, for each kind of work, from our own benchmarks and licensed sources. Further right is better; the better rank is in green.
Every shared result
Each model's best result on every measure both have, grouped by benchmark, with the setting that produced it. Under each measure, every model that benchmark has measured is a grey tick, best to the right, so you can see where the two sit in the field. The headline measures of each benchmark come first; the rest are under "Show all". Green marks the better value; where a source reports ranges that overlap, neither is marked.
| Measure | GPT-5 | GPT-5.6 Sol |
|---|---|---|
| Arena (formerly LMArena) reported by Arena (formerly LMArena) · 3 results | ||
| Overall · TextArena rating, higher is better 90 · 17 of 387 measured | 1,435high reasoning | 1,484xhigh reasoning |
| Overall · Text factualityArena rating, higher is better 111 · 18 of 163 measured | 1,421high reasoning | 1,476xhigh reasoning |
| Business, management and finance · TextArena rating, higher is better 118 · 15 of 380 measured | 1,413high reasoning | 1,485xhigh reasoning |
| Artificial Analysis reported by Artificial Analysis · 3 results | ||
| Artificial Analysis Intelligence Indexindex score, higher is better 90 · 12 of 249 measured | 23.0high reasoning | 47.0max reasoning |
| GPQA Diamond% of questions, higher is better 78 · 5 of 232 measured | 85.4%high reasoning | 94.1%max reasoning |
| Humanity's Last Exam% of questions, higher is better 87 · 11 of 250 measured | 28.5%high reasoning | 49.5%max reasoning |
| SimpleQA Verified (Epoch AI) reported by SimpleQA Verified (Epoch AI) · 1 result | ||
| Correct answers% of questions, higher is better 24 · 7 of 76 measured | 50.1%high reasoning | 69.7%max reasoning |
| UGI Leaderboard reported by UGI Leaderboard · 3 results | ||
| Writing score · UGIscore out of 100, higher is better 68 · 27 of 230 measured | 53.5high reasoning | 65.4medium reasoning |
| Requested-length error · UGI% off the requested word count, lower is better 207 · 3 of 230 measured | 58.0%high reasoning | 1.0%xhigh reasoning |
| Style adherence · UGIscore from 0 to 1, higher is better 20 · 43 of 230 measured | 0.39high reasoning | 0.37xhigh reasoning |
| Vectara Hallucination Leaderboard reported by Vectara Hallucination Leaderboard · 2 results | ||
| Hallucination rate · Vectara% of summaries, lower is better 91 · 83 of 104 measured | 14.7%minimal reasoning | 12.4% |
| Answer rate · Vectara% of documents, higher is better 17 · 61 of 104 measured | 99.9%high reasoning | 99.1% |
Show all 24 results (12 more)
| Measure | GPT-5 | GPT-5.6 Sol |
|---|---|---|
| Arena (formerly LMArena) reported by Arena (formerly LMArena) · 5 results | ||
| Creative writing · TextArena rating, higher is better 128 · 16 of 385 measured | 1,373high reasoning | 1,468xhigh reasoning |
| Expert prompts · TextArena rating, higher is better 90 · 8 of 338 measured | 1,459high reasoning | 1,535xhigh reasoning |
| Instruction following · TextArena rating, higher is better 110 · 9 of 387 measured | 1,410high reasoning | 1,490xhigh reasoning |
| Overall · SearchArena rating, higher is better 22 · 1 of 33 measured | 1,183arena search | 1,256xhigh reasoning, arena search |
| Writing, literature and language · TextArena rating, higher is better 113 · 9 of 386 measured | 1,396high reasoning | 1,481xhigh reasoning |
| Artificial Analysis reported by Artificial Analysis · 7 results | ||
| Terminal-Bench 2.1% of tasks, higher is better 79 · 3 of 131 measured | 35.2%high reasoning | 89.5%xhigh reasoning |
| τ-bench banking% of tasks, higher is better 45 · 14 of 121 measured | 22.1%high reasoning | 44.3%max reasoning |
| Artificial Analysis Coding Indexindex score, higher is better 75 · 2 of 130 measured | 37.8high reasoning | 78.3xhigh reasoning |
| IFBench% of instructions, higher is better 36 · 38 of 195 measured | 73.1%high reasoning | 72.7%max reasoning |
| Long-context reasoning (AA-LCR)% of questions, higher is better 64 · 7 of 241 measured | 78.2%high reasoning | 84.0%max reasoning |
| Terminal-Bench Hard% of tasks, higher is better 39 · 1 of 190 measured | 37.9%medium reasoning | 65.9%max reasoning |
| τ²-bench telecom% of tasks, higher is better 54 · 59 of 193 measured | 86.6%medium reasoning | 85.1%max reasoning |
Quick answers
Is GPT-5 better than GPT-5.6 Sol?
Pick GPT-5.6 Sol if quality matters most: it scores higher (Intelligence Index 47.0 against 23.0). Pick GPT-5 if cost matters more: it costs 2.3 times less per token.
Which is cheaper, GPT-5 or GPT-5.6 Sol?
GPT-5 is cheaper: $3.44 against $8.00 per million tokens, 2.3 times less.
Which is better for agents and tool use, GPT-5 or GPT-5.6 Sol?
GPT-5.6 Sol ranks 12 of 159 models on sale for multi-step tasks with tools: support desks, coding agents, function calls, against 86 for GPT-5.
Which is better for professional work, GPT-5 or GPT-5.6 Sol?
GPT-5.6 Sol ranks 1 of 51 models on sale for real tasks from banking, consulting and law, business documents and freelance projects, against 44 for GPT-5.
Which is better for reasoning and knowledge, GPT-5 or GPT-5.6 Sol?
GPT-5.6 Sol ranks 10 of 187 models on sale for hard questions across science, maths and general knowledge, against 85 for GPT-5.
Which is better for writing, GPT-5 or GPT-5.6 Sol?
GPT-5.6 Sol ranks 22 of 169 models on sale for what people prefer in blind comparisons, and judged writing quality, against 69 for GPT-5.
Which is better for sticking to the facts, GPT-5 or GPT-5.6 Sol?
GPT-5.6 Sol ranks 45 of 139 models on sale for summarising without inventing things, and factual answers, against 93 for GPT-5.
Which has the bigger context window, GPT-5 or GPT-5.6 Sol?
GPT-5.6 Sol: 1,050,000 tokens, against 400,000 for GPT-5.
Run GPT-5 and GPT-5.6 Sol on your own prompt.
Benchmarks aren't your task. Test both side by side on your prompt, with every answer checked against your rules and the cost of each one.