GPT-6.1 Sol vs o3 (2025-04-16)
Humanity's Last Exam Humanity's Last Exam: Very hard expert questions across many fields, written to stump today’s models
GPT-6.1 Sol49.9
o3 (2025-04-16)—
Terminal-Bench 4.0 Terminal-Bench 4.0 · source setting 2: Agents use a terminal to complete executable coding, system administration and research tasks in the 4.0 suite
GPT-6.1 Sol48
o3 (2025-04-16)—
AutomationBench-AA AutomationBench-AA: Agents orchestrate SaaS workflows through APIs; objectives receive credit only when the task has no guardrail violation
GPT-6.1 Sol62.6
o3 (2025-04-16)—
Arena Text · creative writing Arena Text · creative writing: Human pairwise preferences in the stated text vote pool
GPT-6.1 Sol · max1,458
o3 (2025-04-16)—
MMMU-Pro (vision) MMMU-Pro (vision): College-level questions that require understanding images: charts, diagrams, photos
GPT-6.1 Sol83.9
o3 (2025-04-16)—
| Benchmark | GPT-6.1 Sol | o3 (2025-04-16) |
|---|---|---|
| Humanity's Last Exam | 49.9 | — |
| Terminal-Bench 4.0 | 48 | — |
| AutomationBench-AA | 62.6 | — |
| Arena Text · creative writing | 1,458 · max | — |
| MMMU-Pro (vision) | 83.9 | — |
Specs
| Spec | GPT-6.1 Sol | o3 (2025-04-16) |
|---|---|---|
| Input, per 1M | $2 | !$2 |
| Output, per 1M | $10 | !$8 |
| Cost per task? | $0.21 | — |
| Speed, tok/s | 45 | — |
| Context | 1.05M | 200K |
| Weights | Closed | Not verified |
| Released | Sep 2026 | Apr 2025 |
| See images | ✓ | ✓ |
| Read PDFs | Partial | ? |
| Hear audio | — | — |
| Watch video | — | — |
| Call tools | ✓ | ✓ |
| Force a tool call | Partial | ? |
| Return strict JSON | ✓ | ✓ |
| Search the web | ✓ | ? |
| Run code | ✓ | ? |
| Use a computer | ✓ | ? |
What breaks when you switch
GPT-6.1 Sol → o3 (2025-04-16)
Context: 1.05M → 200K
o3 (2025-04-16) → GPT-6.1 Sol
Remove unsupported parameters›
With reasoning enabled, remove temperature, top_p and top_logprobs, plus Chat Completions logprobs or Responses message.output_text.logprobs includes.
Responses API for tool calling · Chat Completions›
Use Responses for tools; Chat Completions supports this model without tool calling.
No none or minimal reasoning›
The none and minimal reasoning efforts are unsupported; choose low, medium, high, xhigh or max.
More than 272K input tokens›
Prompts above 272K input tokens charge 2× input and cache rates and 1.5× output rates for the entire request.
Cache writes billed at 1.25x›
Cache writes cost 1.25× the uncached input token rate.
Ultrafast prices 6x Standard›
Ultrafast mode costs 6× Standard token prices.