Introducing Parallel Search Turbo: the fastest and lowest-cost search for AI. [Learn more.](/content/blog/parallel-search-turbo/index.html)[\[Learn more.\](/content/blog/parallel-search-turbo/index.html)](/content/ai/blog/parallel-search-turbo/index.html)

HumanMachine

April 7, 2026

# \\# A new deep research frontier on DeepSearchQA with the Task API Harness

Tags:[Benchmarks](/content/blog?tag=benchmarks/index.html)

Reading time: 7 min

The Parallel Task API is the most powerful deep research agent on the market. It’s used in production agents by teams at Opendoor, Attio, Modal, Starbridge, Profound, and more.

Parallel’s [Processor](https://docs.parallel.ai/task-api/guides/choose-a-processor)[\[Processor\](https://docs.parallel.ai/task-api/guides/choose-a-processor)](https://docs.parallel.ai/task-api/guides/choose-a-processor) architecture allows for flexibility and fine-tuned application of different compute budgets depending on the complexity of a research task. The “Ultra” range of Task API Processors is state-of-the-art on DeepSearchQA:

- \- Parallel Ultra is 11% more accurate and up to 57% lower cost versus the next best, GPT-5.4.
- \- With higher compute budgets, Parallel Ultra2x, 4x, and 8x continue to push the Pareto frontier of accuracy and cost, achieving 82% accuracy at the highest end.

This blog details some of the techniques we’ve employed to achieve this state of the art research quality.

## \\#\# Results

We evaluated Parallel's Task API Processors against GPT 5.4, Opus 4-6, Gemini 3.1 Pro, Exa Search Deep Reasoning, and Perplexity Sonar Pro.

!\[\](https://cdn.sanity.io/images/5hzduz3y/production/a24599e688575e2f45db79f7d4063f64688f8570-4800x2700.jpg)

| Provider | Model | Cost (CPM) | Accuracy (%) |
| --- | --- | --- | --- |
| Parallel | Ultra 8x | $2400 | 82 |
| Parallel | Ultra 4x | $1200 | 81 |
| Parallel | Ultra 2x | $600 | 77 |
| Parallel | Ultra | $300 | 70 |
| OpenAI | GPT 5.4 with code execution | $701 | 63 |
| Google | Gemini 3.1 Pro with code execution | $707 | 62 |
| Anthropic | Opus 4-6 with PTC | $36,231\* | 58 |
| Perplexity | Sonar Pro | $883 | 28 |
| Exa | Search Deep Reasoning | $15 | 18 |

_\_CPM: USD per 1,000 requests. Cost is shown on a log scale.\__

_\_\*The cost of Opus 4-6 is higher than expected due to Anthropic’s potential billing issue, where prompt caching savings for PTC are not passed on to the user.\__

## \\#\# About DeepSearchQA

DeepSearchQA is a 900-question evaluation from Google designed to test agents on multi-step information-seeking tasks across 17 fields of expertise. Each question is a causal chain: you can't answer the second part without resolving the first, and you can't resolve the first without searching the web, reading the results, and reasoning about what to search next.

A typical query might ask: _\_identify every researcher who co-authored papers with a specific professor at three different institutions over a decade, then determine which of those co-authors later joined a federal advisory committee.\__ Simple web retrieval won't cut it. The agent needs to plan a research strategy, execute multiple searches, cross-reference results across sources, and synthesize a precise answer with zero false positives.

Here, accuracy on DeepSearchQA means "fully correct": the response must be semantically identical to the ground-truth set. The agent has to find all correct answers while including none that are wrong.

_\_\*We use DeepSearchQA instead of BrowseComp to ensure more reliable evaluation, as some models have begun to memorize portions of the BrowseComp dataset, potentially inflating performance.\__

## \\#\# Inside the Task API Harness

!\[\](https://cdn.sanity.io/images/5hzduz3y/production/af4801e37c89acb2b3dfa0002ef6ead13dea4c4c-900x600.gif)Visualization of the Parallel Task API Harness

Our results stem from combining several techniques.

- \- **\*\*Code Execution: \*\*** We enable programmatic tool use instead of relying purely on text-based interactions. This allows for more robust state and context management, and prevents unnecessary context growth by keeping intermediate reasoning and tool outputs externalized.
- \- **\*\*Aggressive prompt caching: \*\*** We cache repeated prompts and intermediate results wherever possible. This is critical for both latency and cost efficiency at scale.
- \- **\*\*Budget-aware execution: \*\*** Our system dynamically adapts its behavior based on a target budget, allocating resources where they have the highest impact on quality. This ensures we consistently achieve the best possible performance for a given cost envelope.
- \- **\*\*Context compaction: \*\*** As context grows, we proactively compress and distill it to retain only the most relevant information. This helps maintain model performance while avoiding degradation from overly long contexts.
- \- **\*\*Search and extraction infrastructure: \*\*** Our Task API is built over Parallel’s proprietary Search and Extract APIs. These APIs are built from the ground up to optimize for agentic workloads, enabling higher recall and more precise retrieval across heterogeneous sources. This ensures the model operates over clean and relevant inputs, instead of noisy or redundant inputs.

Together, these techniques allow us to scale reasoning depth and reliability while maintaining efficiency, leading to state-of-the-art results on DeepSearchQA. Instead of having the model orchestrate tools through tool calling, we gave it the ability to write and execute code.

The orchestrating model generates Python that calls research tools as ordinary functions. This code runs in a sandboxed interpreter. Only the final output of each code block re-enters the model's context. Intermediate data stays in the interpreter's variable state, not in the conversation history.

!\[\](https://cdn.sanity.io/images/5hzduz3y/production/dfb06ffe55ea24e567f370fa1ce0b26be8b4d772-1920x1004.png)Task API high-level architecture

````
1
2
3
4
5

Step 1: search("company X revenue 2024") → [5 results in context]
Step 2: extract(url_1) → [full page content in context]
Step 3: extract(url_2) → [full page content in context]
Step 4: search("company X revenue 2023") → [5 more results in context]
...context grows with every step``` Step 1: search("company X revenue 2024") → [5 results in context]
Step 2: extract(url_1) → [full page content in context]
Step 3: extract(url_2) → [full page content in context]
Step 4: search("company X revenue 2023") → [5 more results in context]
...context grows with every step
```
````

Each step inflates the context window. By step 8, the model spends most of its capacity re-reading old results.

````
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19

# First, use one search step to discover the URL pattern.
results_2024 = search("Company X 2024 annual report")
report_2024 = results_2024[0].url
# e.g. https://investors.companyx.com/financials/annual-reports/2024-annual-report.pdf

# Infer a reusable template from the 2024 URL and a parse function parse_fn.
url_template = "https://investors.companyx.com/financials/annual-reports/{year}-annual-report.pdf"

years = [2022, 2023, 2024]
reports = {}

for year in years:
    url = url_template.format(year=year)
    reports[year] = parse_fn(extract(
        question=f"What was Company X's reported revenue in {year}? Return the value and supporting quote.",
        urls=[url],
    ))

return reports``` # First, use one search step to discover the URL pattern.
results_2024 = search("Company X 2024 annual report")
report_2024 = results_2024[0].url
# e.g. https://investors.companyx.com/financials/annual-reports/2024-annual-report.pdf

years = [2022, 2023, 2024]
reports = {}

return reports
```
````

Multiple searches, extractions, and analyses happen in a single execution step. The full page content of both extracted pages (potentially tens of thousands of tokens) never enters the research agent's context. Only the extracted revenue figures flow back.

This compounds. A 20-step research task that would fill a 128K context window under tool calling stays under 30K tokens because intermediate data lives in the interpreter, not the conversation.

### \\#\#\# Persistent state as working memory

Variables created in one code execution step survive to the next. When the model writes _\_findings\["report"\] = report\_2024\__ in iteration 3, that variable is still accessible in iteration 7 when the model needs to cross-reference revenue against employee headcount.

This creates a separation that matters: the conversation history captures the model's high-level reasoning trajectory, while the interpreter's variable state captures the raw data it has gathered. The conversation can be compacted without losing granular data.

### \\#\#\# Inside the sandbox

Running LLM-generated code in production requires strong isolation. Our interpreter is a sandboxed Python runtime built on Rust with no access to the network, filesystem, or operating system. The only way code interacts with the outside world is through explicitly injected functions: search, extract and a handful of state management utilities.

!\[\](https://cdn.sanity.io/images/5hzduz3y/production/614565599ea2bfe5f962306c615fca92bc6e82a5-2080x1144.png)Sandbox architecture

This boundary gives us two things. First, the model can write complex data processing logic (filtering, aggregation, string manipulation, conditional branching) without risk of side effects.

### \\#\#\# Budget-aware execution

One of the attractive features of Parallel’s Task API is fixed and predictable costs rather than per-token pricing. We have built our agent harness to support budgets as a first-class concept. This budget isn’t just a fixed limit on the number of steps. A fixed step limit penalizes simple and complex queries equally. We track cumulative cost across iterations instead. The system monitors token spend across all LLM calls, both the orchestrating model and any sub-model invocations, and injects budget warnings when remaining spend drops below a threshold. When the budget is nearly exhausted, the model synthesizes its findings and produces a final answer with whatever it has gathered.

A simple factual query might terminate in 2 iterations and cost a few cents. A complex multi-source comparison might run for 15 iterations across a larger budget. The architecture adapts to the difficulty of the question rather than imposing a uniform ceiling. This is also how we offer the full range of Processors from Ultra ($300 CPM) through Ultra8x ($2,400 CPM): higher-tier Processors get a larger budget, which lets the agent pursue more research paths before synthesizing.

### \\#\#\# Context compaction

Even with code execution keeping intermediate data out of context, long research sessions accumulate history: the model's reasoning, code blocks, and execution summaries. When this history approaches context limits, we trigger compaction, a summarization pass that condenses earlier conversation turns while preserving key findings and the current research trajectory.

The persistent variable state in the interpreter is unaffected by compaction (it lives in the interpreter, not the conversation). This means the model can sustain research across many more iterations than its raw context window would suggest.

## \\#\# Why we picked this architecture

Most deep research systems follow a familiar loop: an LLM generates a plan, calls a tool, reads the result, and repeats. This works for simple queries but degrades as research complexity grows. We evaluated several architectural patterns for complex research workflows, each with different trade-offs across context efficiency, granularity of information access, and adaptability.

| Architecture | Cost efficiency | Fine-grained Research Control | Adaptability | Reason |
| --- | --- | --- | --- | --- |
| Naive Agent Loop (LLM → tool → result → LLM → …) | ❌ | ❌ | ✅ | Maximally flexible, but context grows at every step. The model ends up spending too much capacity re-reading intermediate results. |
| Agent Loop with context compression | ✅ | ❌ | 🟠 | Keeps context manageable, but important details are often compressed away. This makes it harder to revisit evidence or pursue subtle lines of inquiry. |
| Static Plan + sub-agents | ✅ | 🟠 | ❌ | Works well for predictable workflows, but cannot adapt cleanly once the plan is set. |
| Agent Loop with sub-agents | ✅ | 🟠 | 🟠 | Better runtime flexibility than static planning, but still suffers from coordination and memory fragmentation across agents. |
| Task API Harness (Parallel) | ✅ | ✅ | ✅ | Preserves detailed evidence access while keeping the intermediate state out of the model context. It can adapt mid-run, branch when needed, and scale without bloating the prompt. |

````
1
2
3
4
5
6
7
8
9
10
11
12
13
14

import parallel

client = parallel.Client(api_key="your-api-key")

task = client.task_runs.create(
    objective="Identify every researcher who co-authored papers "
              "with Dr. Maria Chen at Stanford, MIT, and Caltech "
              "between 2010 and 2020, then determine which of those "
              "co-authors later joined the NIH Advisory Committee "
              "to the Director.",
    processor="ultra8x",
)

print(task.output)``` import parallel

client = parallel.Client(api_key="your-api-key")

print(task.output)
```
````

## \\#\# About the Parallel Task API

The Task API is a general-purpose web research agent API. Define what you need in natural language or structured JSON, and it handles research, synthesis, and structured output with citations and confidence levels. Processors range from Lite (basic lookups, $5/1K) through Ultra8x (the hardest deep research, $2,400/1K).

## \\#\# About Parallel Web Systems

Parallel builds web infrastructure for AI. Our APIs, including Search, Extract, Task, FindAll, and Monitor, give AI agents structured, grounded access to the open web, powered by a rapidly growing proprietary index of the global internet.

Parallel turns human workflows that took days into agentic workflows that take seconds. Fortune 100s and leading frontier AI companies including Harvey, Manus, Starbridge, and Profound rely on Parallel for legal grounding, fact-checking, contract monitoring, and high-quality content generation.

Get started at [platform.parallel.ai](http://platform.parallel.ai/)[\[platform.parallel.ai\](http://platform.parallel.ai)](http://platform.parallel.ai/)

## \\#\# Ready to get started?

Sign up for free. No credit card required.

[Try Parallel](https://platform.parallel.ai/home)[\[Try Parallel\](https://platform.parallel.ai/home)](https://platform.parallel.ai/home) [Contact sales](https://contact.parallel.ai/)[\[Contact sales\](https://contact.parallel.ai/)](https://contact.parallel.ai/)

[Are you an agent? Read this to onboard Parallel](/content/agents.md)[\[Are you an agent? Read this to onboard Parallel\](/content/agents.md)](/content/ai/agents.md)

By Parallel

April 7, 2026

### \\#\# Related Posts80

[Jul 30, 2026\\
**\- \[Building an always-on background agent to proactively support customers\](/content/blog/customer-watch-background-agent/index.html)**](/content/blog/customer-watch-background-agent/index.html)

[Tags:](/content/blog/customer-watch-background-agent/index.html) [Developers](/content/blog?tag=developers/index.html)

Author: By Khushi Shelat

[Jul 21, 2026\\
**\- \[Introducing the Parallel Responses API\](/content/blog/responses-api/index.html)**](/content/blog/responses-api/index.html)

[Tags:](/content/blog/responses-api/index.html) [Product](/content/blog?tag=product/index.html)

Author: By Parallel

[Jul 20, 2026\\
**\- \[Building a vendor intelligence system with Parallel\](/content/blog/vendor-intelligence-system/index.html)**](/content/blog/vendor-intelligence-system/index.html)

[Tags:](/content/blog/vendor-intelligence-system/index.html) [Developers](/content/blog?tag=developers/index.html)

Author: By Sahith Jagarlamudi

[Jul 16, 2026\\
**\- \[Parallel and Google Cloud Announce Partnership for Agentic Web Search on Gemini Enterprise Agent Platform\](/content/blog/google-cloud-partnership/index.html)**](/content/blog/google-cloud-partnership/index.html)

[Tags:](/content/blog/google-cloud-partnership/index.html) [Product](/content/blog?tag=product/index.html)

Author: By Parallel

[Jul 15, 2026\\
**\- \[$5 in free Parallel credits, every month\](/content/blog/free-tier-parallel/index.html)**](/content/blog/free-tier-parallel/index.html)

[Tags:](/content/blog/free-tier-parallel/index.html) [Product](/content/blog?tag=product/index.html)

Author: By Parallel

\\
\\
Jul 13, 2026\\
\\
**\- \[Introducing Parallel Search Turbo\](/content/blog/parallel-search-turbo/index.html)** \\
\\
Author: By Parallel](/content/blog/parallel-search-turbo/index.html)

[Jul 12, 2026\\
**\- \[Building a realtime voice agent with GPT-Realtime-2.1 and Parallel Search Turbo\](/content/blog/gpt-realtime-parallel-turbo/index.html)**](/content/blog/gpt-realtime-parallel-turbo/index.html)

[Tags:](/content/blog/gpt-realtime-parallel-turbo/index.html) [Developers](/content/blog?tag=developers/index.html)

Author: By George Pickett

[Jul 10, 2026\\
**\- \[How Nooks cut web search costs 70.5% by switching to Parallel\](/content/blog/case-study-nooks/index.html)**](/content/blog/case-study-nooks/index.html)

[Tags:](/content/blog/case-study-nooks/index.html) [Customers](/content/blog?tag=customers/index.html)

Author: By Parallel

[Jul 8, 2026\\
**\- \[How Build created live geofenced alerts powered by Parallel for institutional real estate\](/content/blog/case-study-build/index.html)**](/content/blog/case-study-build/index.html)

[Tags:](/content/blog/case-study-build/index.html) [Customers](/content/blog?tag=customers/index.html)

Author: By Parallel

[Jun 9, 2026\\
**\- \[OpenClaw now has free, LLM-optimized web search by default powered by Parallel\](/content/blog/free-web-search-openclaw/index.html)**](/content/blog/free-web-search-openclaw/index.html)

[Tags:](/content/blog/free-web-search-openclaw/index.html) [Company](/content/blog?tag=company/index.html)

Author: By Parallel

[Jun 5, 2026\\
**\- \[Introducing real-time Entity Search\](/content/blog/entity-search-company/index.html)**](/content/blog/entity-search-company/index.html)

[Tags:](/content/blog/entity-search-company/index.html) [Product](/content/blog?tag=product/index.html)

Author: By Parallel

[Jun 3, 2026\\
**\- \[How we enrich & triage inbound leads using the Parallel Task API\](/content/blog/enrich-triage-inbound-leads-parallel-task-api/index.html)**](/content/blog/enrich-triage-inbound-leads-parallel-task-api/index.html)

[Tags:](/content/blog/enrich-triage-inbound-leads-parallel-task-api/index.html) [Developers](/content/blog?tag=developers/index.html)

Author: By Khushi Shelat

[May 20, 2026\\
**\- \[How AirOps creates citation-worthy content at scale, powered by Parallel\](/content/blog/case-study-airops/index.html)**](/content/blog/case-study-airops/index.html)

[Tags:](/content/blog/case-study-airops/index.html) [Customers](/content/blog?tag=customers/index.html)

Author: By Parallel

[May 18, 2026\\
**\- \[Introducing Index by Parallel\](/content/blog/introducing-index-by-parallel/index.html)**](/content/blog/introducing-index-by-parallel/index.html)

[Tags:](/content/blog/introducing-index-by-parallel/index.html) [Product](/content/blog?tag=product/index.html)

Author: By Parallel

[May 7, 2026\\
**\- \[Parallel Monitor API: New processor tiers, snapshots and event streams, and Basis on every event\](/content/blog/monitor-api/index.html)**](/content/blog/monitor-api/index.html)

[Tags:](/content/blog/monitor-api/index.html) [Product](/content/blog?tag=product/index.html)

Author: By Parallel

[May 4, 2026\\
**\- \[How we built parallelmpp.dev\](/content/blog/parallel-mpp-dev/index.html)**](/content/blog/parallel-mpp-dev/index.html)

[Tags:](/content/blog/parallel-mpp-dev/index.html) [Developers](/content/blog?tag=developers/index.html)

Author: By Son Do

[Apr 29, 2026\\
**\- \[How Actively's Per Account Agents use Parallel to turn the entire web into a proactive sales intelligence layer\](/content/blog/case-study-actively/index.html)**](/content/blog/case-study-actively/index.html)

[Tags:](/content/blog/case-study-actively/index.html) [Customers](/content/blog?tag=customers/index.html)

Author: By Parallel

[Apr 28, 2026\\
**\- \[Parallel Raises at $2 Billion Valuation to Scale Web Infrastructure for Agents\](/content/blog/series-b/index.html)**](/content/blog/series-b/index.html)

[Tags:](/content/blog/series-b/index.html) [Company](/content/blog?tag=company/index.html)

Author: By Parallel

[Apr 24, 2026\\
**\- \[Building a free CLI agent with Pi, Ollama, Gemma 4, and Parallel\](/content/blog/free-CLI-agent/index.html)**](/content/blog/free-CLI-agent/index.html)

[Tags:](/content/blog/free-CLI-agent/index.html) [Developers](/content/blog?tag=developers/index.html)

Author: By Matt Harris

[Apr 23, 2026\\
**\- \[Parallel Search is now free for agents via MCP\](/content/blog/free-web-search-mcp/index.html)**](/content/blog/free-web-search-mcp/index.html)

[Tags:](/content/blog/free-web-search-mcp/index.html) [Product](/content/blog?tag=product/index.html)

Author: By Parallel

[Apr 21, 2026\\
**\- \[Upgrades to the Parallel Search & Extract APIs\](/content/blog/parallel-search-api/index.html)**](/content/blog/parallel-search-api/index.html)

[Tags:](/content/blog/parallel-search-api/index.html) [Benchmarks](/content/blog?tag=benchmarks/index.html)

Author: By Parallel

[Apr 20, 2026\\
**\- \[How Finch is scaling plaintiff law with AI agents that research like associates\](/content/blog/case-study-finch/index.html)**](/content/blog/case-study-finch/index.html)

[Tags:](/content/blog/case-study-finch/index.html) [Customers](/content/blog?tag=customers/index.html)

Author: By Parallel

[Apr 8, 2026\\
**\- \[Genpact and Parallel Web Systems Partner to Drive Tangible Efficiency from AI Systems\](/content/blog/genpact-parallel-partnership/index.html)**](/content/blog/genpact-parallel-partnership/index.html)

[Tags:](/content/blog/genpact-parallel-partnership/index.html) [Company](/content/blog?tag=company/index.html)

Author: By Parallel

[Apr 8, 2026\\
**\- \[How Genpact helps top US insurers cut contents claims processing times in half with Parallel \](/content/blog/case-study-genpact/index.html)**](/content/blog/case-study-genpact/index.html)

[Tags:](/content/blog/case-study-genpact/index.html) [Customers](/content/blog?tag=customers/index.html)

Author: By Parallel

[Mar 30, 2026\\
**\- \[How Modal saves tens of thousands annually by building in-house GTM pipelines with Parallel\](/content/blog/case-study-modal/index.html)**](/content/blog/case-study-modal/index.html)

[Tags:](/content/blog/case-study-modal/index.html) [Customers](/content/blog?tag=customers/index.html)

Author: By Parallel

[Mar 25, 2026\\
**\- \[How Opendoor uses Parallel as the enterprise grade web research layer powering its AI-native real estate operations\](/content/blog/case-study-opendoor/index.html)**](/content/blog/case-study-opendoor/index.html)

[Tags:](/content/blog/case-study-opendoor/index.html) [Customers](/content/blog?tag=customers/index.html)

Author: By Parallel

[Mar 19, 2026\\
**\- \[Introducing stateful web research agents with multi-turn conversations\](/content/blog/task-api-interactions/index.html)**](/content/blog/task-api-interactions/index.html)

[Tags:](/content/blog/task-api-interactions/index.html) [Product](/content/blog?tag=product/index.html)

Author: By Parallel

[Mar 18, 2026\\
**\- \[Parallel is live on Tempo, now available natively to agents with the Machine Payments Protocol\](/content/blog/tempo-stripe-mpp/index.html)**](/content/blog/tempo-stripe-mpp/index.html)

[Tags:](/content/blog/tempo-stripe-mpp/index.html) [Company](/content/blog?tag=company/index.html)

Author: By Parallel

[Mar 17, 2026\\
**\- \[How Parallel helped Kepler build AI that finance professionals can actually trust\](/content/blog/case-study-kepler/index.html)**](/content/blog/case-study-kepler/index.html)

[Tags:](/content/blog/case-study-kepler/index.html) [Customers](/content/blog?tag=customers/index.html)

Author: By Parallel

[Mar 10, 2026\\
**\- \[Introducing the Parallel CLI\](/content/blog/parallel-cli/index.html)**](/content/blog/parallel-cli/index.html)

[Tags:](/content/blog/parallel-cli/index.html) [Product](/content/blog?tag=product/index.html)

Author: By Parallel

[Mar 4, 2026\\
**\- \[How Profound helps brands win AI Search with high-quality web research and content creation powered by Parallel\](/content/blog/case-study-profound/index.html)**](/content/blog/case-study-profound/index.html)

[Tags:](/content/blog/case-study-profound/index.html) [Customers](/content/blog?tag=customers/index.html)

Author: By Parallel

[Mar 2, 2026\\
**\- \[How Harvey is expanding legal AI internationally with Parallel\](/content/blog/case-study-harvey/index.html)**](/content/blog/case-study-harvey/index.html)

[Tags:](/content/blog/case-study-harvey/index.html) [Customers](/content/blog?tag=customers/index.html)

Author: By Parallel

[Feb 23, 2026\\
**\- \[How Tabstack by Mozilla enables agents to navigate the web with Parallel’s best-in-class web search\](/content/blog/case-study-tabstack/index.html)**](/content/blog/case-study-tabstack/index.html)

[Tags:](/content/blog/case-study-tabstack/index.html) [Customers](/content/blog?tag=customers/index.html)

Author: By Parallel

[Feb 4, 2026\\
**\- \[Parallel Web Tools and Agents now available across Vercel AI Gateway, AI SDK, and Marketplace\](/content/blog/vercel/index.html)**](/content/blog/vercel/index.html)

[Tags:](/content/blog/vercel/index.html) [Product](/content/blog?tag=product/index.html)

Author: By Parallel

[Jan 28, 2026\\
**\- \[Authenticated page access for the Parallel Task API\](/content/blog/authenticated-page-access/index.html)**](/content/blog/authenticated-page-access/index.html)

[Tags:](/content/blog/authenticated-page-access/index.html) [Product](/content/blog?tag=product/index.html)

Author: By Parallel

[Jan 21, 2026\\
**\- \[Introducing structured outputs for the Monitor API\](/content/blog/structured-outputs-monitor/index.html)**](/content/blog/structured-outputs-monitor/index.html)

[Tags:](/content/blog/structured-outputs-monitor/index.html) [Product](/content/blog?tag=product/index.html)

Author: By Parallel

[Jan 15, 2026\\
**\- \[Introducing research models with Basis for the Parallel Chat API\](/content/blog/research-models-chat/index.html)**](/content/blog/research-models-chat/index.html)

[Tags:](/content/blog/research-models-chat/index.html) [Product](/content/blog?tag=product/index.html)

Author: By Parallel

[Jan 8, 2026\\
**\- \[Build a real-time fact checker with Parallel and Cerebras\](/content/blog/cerebras-fact-checker/index.html)**](/content/blog/cerebras-fact-checker/index.html)

[Tags:](/content/blog/cerebras-fact-checker/index.html) [Developers](/content/blog?tag=developers/index.html)

Author: By Parallel

[Dec 17, 2025\\
**\- \[Parallel Task API achieves state-of-the-art accuracy on DeepSearchQA\](/content/blog/deepsearch-qa/index.html)**](/content/blog/deepsearch-qa/index.html)

[Tags:](/content/blog/deepsearch-qa/index.html) [Benchmarks](/content/blog?tag=benchmarks/index.html)

Author: By Parallel

[Dec 16, 2025\\
**\- \[Introducing Granular Basis for the Task API\](/content/blog/granular-basis-task-api/index.html)**](/content/blog/granular-basis-task-api/index.html)

[Tags:](/content/blog/granular-basis-task-api/index.html) [Product](/content/blog?tag=product/index.html)

Author: By Parallel

[Dec 11, 2025\\
**\- \[How Amp’s coding agents build better software with Parallel Search\](/content/blog/case-study-amp/index.html)**](/content/blog/case-study-amp/index.html)

[Tags:](/content/blog/case-study-amp/index.html) [Customers](/content/blog?tag=customers/index.html)

Author: By Parallel

[Dec 10, 2025\\
**\- \[Latency improvements on the Parallel Task API \](/content/blog/task-api-latency/index.html)**](/content/blog/task-api-latency/index.html)

[Tags:](/content/blog/task-api-latency/index.html) [Product](/content/blog?tag=product/index.html)

Author: By Parallel

[Nov 20, 2025\\
**\- \[Introducing Parallel Extract\](/content/blog/introducing-parallel-extract/index.html)**](/content/blog/introducing-parallel-extract/index.html)

[Tags:](/content/blog/introducing-parallel-extract/index.html) [Product](/content/blog?tag=product/index.html)

Author: By Parallel

[Nov 18, 2025\\
**\- \[Introducing Parallel FindAll\](/content/blog/introducing-findall-api/index.html)**](/content/blog/introducing-findall-api/index.html)

[Tags:](/content/blog/introducing-findall-api/index.html) [Product](/content/blog?tag=product/index.html),[Benchmarks](/content/blog?tag=benchmarks/index.html)

Author: By Parallel

[Nov 13, 2025\\
**\- \[Introducing Parallel Monitor\](/content/blog/monitor-api-beta/index.html)**](/content/blog/monitor-api-beta/index.html)

[Tags:](/content/blog/monitor-api-beta/index.html) [Product](/content/blog?tag=product/index.html)

Author: By Parallel

[Nov 12, 2025\\
**\- \[Parallel raises $100M Series A to build web infrastructure for agents\](/content/blog/series-a/index.html)**](/content/blog/series-a/index.html)

[Tags:](/content/blog/series-a/index.html) [Company](/content/blog?tag=company/index.html)

Author: By Parallel

[Nov 11, 2025\\
**\- \[How Macroscope reduced code review false positives with Parallel\](/content/blog/case-study-macroscope/index.html)**](/content/blog/case-study-macroscope/index.html)

[Tags:](/content/blog/case-study-macroscope/index.html) [Customers](/content/blog?tag=customers/index.html)

Author: By Parallel

[Nov 6, 2025\\
**\- \[Introducing Parallel Search\](/content/blog/parallel-search-api-beta/index.html)**](/content/blog/parallel-search-api-beta/index.html)

[Tags:](/content/blog/parallel-search-api-beta/index.html) [Benchmarks](/content/blog?tag=benchmarks/index.html)

Author: By Parallel

[Nov 3, 2025\\
**\- \[Parallel processors set new price-performance standard on SealQA benchmark\](/content/blog/benchmarks-task-api-sealqa/index.html)**](/content/blog/benchmarks-task-api-sealqa/index.html)

[Tags:](/content/blog/benchmarks-task-api-sealqa/index.html) [Benchmarks](/content/blog?tag=benchmarks/index.html)

Author: By Parallel

[Oct 30, 2025\\
**\- \[Introducing LLMTEXT, an open source toolkit for the llms.txt standard\](/content/blog/LLMTEXT-for-llmstxt/index.html)**](/content/blog/LLMTEXT-for-llmstxt/index.html)

[Tags:](/content/blog/LLMTEXT-for-llmstxt/index.html) [Product](/content/blog?tag=product/index.html)

Author: By Parallel

[Oct 23, 2025\\
**\- \[How Starbridge powers public sector GTM with state-of-the-art web research\](/content/blog/case-study-starbridge/index.html)**](/content/blog/case-study-starbridge/index.html)

[Tags:](/content/blog/case-study-starbridge/index.html) [Customers](/content/blog?tag=customers/index.html)

Author: By Parallel

[Oct 22, 2025\\
**\- \[Building a market research platform with Parallel Deep Research\](/content/blog/cookbook-market-research-platform-with-parallel/index.html)**](/content/blog/cookbook-market-research-platform-with-parallel/index.html)

[Tags:](/content/blog/cookbook-market-research-platform-with-parallel/index.html) [Developers](/content/blog?tag=developers/index.html)

Author: By Parallel

[Oct 17, 2025\\
**\- \[How Lindy brings state-of-the-art web research to automation flows\](/content/blog/case-study-lindy/index.html)**](/content/blog/case-study-lindy/index.html)

[Tags:](/content/blog/case-study-lindy/index.html) [Customers](/content/blog?tag=customers/index.html)

Author: By Parallel

[Oct 16, 2025\\
**\- \[Introducing the Parallel Task MCP Server\](/content/blog/parallel-task-mcp-server/index.html)**](/content/blog/parallel-task-mcp-server/index.html)

[Tags:](/content/blog/parallel-task-mcp-server/index.html) [Product](/content/blog?tag=product/index.html)

Author: By Parallel

[Oct 9, 2025\\
**\- \[Introducing the Core2x Processor for improved compute control on the Task API\](/content/blog/core2x-processor/index.html)**](/content/blog/core2x-processor/index.html)

[Tags:](/content/blog/core2x-processor/index.html) [Product](/content/blog?tag=product/index.html)

Author: By Parallel

[Oct 8, 2025\\
**\- \[How Day AI merges private and public data for business intelligence\](/content/blog/case-study-day-ai/index.html)**](/content/blog/case-study-day-ai/index.html)

[Tags:](/content/blog/case-study-day-ai/index.html) [Customers](/content/blog?tag=customers/index.html)

Author: By Parallel

[Oct 7, 2025\\
**\- \[Full Basis framework for all Task API Processors\](/content/blog/full-basis-framework-for-task-api/index.html)**](/content/blog/full-basis-framework-for-task-api/index.html)

[Tags:](/content/blog/full-basis-framework-for-task-api/index.html) [Product](/content/blog?tag=product/index.html)

Author: By Parallel

[Oct 6, 2025\\
**\- \[Building a real-time streaming task manager with Parallel\](/content/blog/cookbook-sse-task-manager-with-parallel/index.html)**](/content/blog/cookbook-sse-task-manager-with-parallel/index.html)

[Tags:](/content/blog/cookbook-sse-task-manager-with-parallel/index.html) [Developers](/content/blog?tag=developers/index.html)

Author: By Parallel

[Sep 30, 2025\\
**\- \[How Gumloop built a new AI automation framework with web intelligence as a core node\](/content/blog/case-study-gumloop/index.html)**](/content/blog/case-study-gumloop/index.html)

[Tags:](/content/blog/case-study-gumloop/index.html) [Customers](/content/blog?tag=customers/index.html)

Author: By Parallel

[Sep 16, 2025\\
**\- \[Introducing the TypeScript SDK\](/content/blog/typescript-sdk/index.html)**](/content/blog/typescript-sdk/index.html)

[Tags:](/content/blog/typescript-sdk/index.html) [Product](/content/blog?tag=product/index.html)

Author: By Parallel

[Sep 12, 2025\\
**\- \[Building a serverless competitive intelligence platform with MCP + Task API\](/content/blog/cookbook-competitor-research-with-reddit-mcp/index.html)**](/content/blog/cookbook-competitor-research-with-reddit-mcp/index.html)

[Tags:](/content/blog/cookbook-competitor-research-with-reddit-mcp/index.html) [Developers](/content/blog?tag=developers/index.html)

Author: By Parallel

[Sep 11, 2025\\
**\- \[Introducing Parallel Deep Research reports\](/content/blog/deep-research-reports/index.html)**](/content/blog/deep-research-reports/index.html)

[Tags:](/content/blog/deep-research-reports/index.html) [Product](/content/blog?tag=product/index.html)

Author: By Parallel

[Sep 9, 2025\\
**\- \[A new pareto-frontier for Deep Research price-performance\](/content/blog/deep-research-benchmarks/index.html)**](/content/blog/deep-research-benchmarks/index.html)

[Tags:](/content/blog/deep-research-benchmarks/index.html) [Benchmarks](/content/blog?tag=benchmarks/index.html)

Author: By Parallel

[Sep 5, 2025\\
**\- \[Building a Full-Stack Search Agent with Parallel and Cerebras\](/content/blog/cookbook-search-agent/index.html)**](/content/blog/cookbook-search-agent/index.html)

[Tags:](/content/blog/cookbook-search-agent/index.html) [Developers](/content/blog?tag=developers/index.html)

Author: By Parallel

[Aug 21, 2025\\
**\- \[Webhooks for the Parallel Task API\](/content/blog/webhooks/index.html)**](/content/blog/webhooks/index.html)

[Tags:](/content/blog/webhooks/index.html) [Product](/content/blog?tag=product/index.html)

Author: By Parallel

[Aug 14, 2025\\
**\- \[Introducing Parallel: Web Search Infrastructure for AIs \](/content/blog/introducing-parallel/index.html)**](/content/blog/introducing-parallel/index.html)

[Tags:](/content/blog/introducing-parallel/index.html) [Benchmarks](/content/blog?tag=benchmarks/index.html),[Product](/content/blog?tag=product/index.html)

Author: By Parallel

[Aug 7, 2025\\
**\- \[Introducing SSE for Task Runs\](/content/blog/sse-for-tasks/index.html)**](/content/blog/sse-for-tasks/index.html)

[Tags:](/content/blog/sse-for-tasks/index.html) [Product](/content/blog?tag=product/index.html)

Author: By Parallel

[Aug 5, 2025\\
**\- \[A new line of advanced Processors: Ultra2x, Ultra4x, and Ultra8x \](/content/blog/new-advanced-processors/index.html)**](/content/blog/new-advanced-processors/index.html)

[Tags:](/content/blog/new-advanced-processors/index.html) [Product](/content/blog?tag=product/index.html)

Author: By Parallel

[Aug 4, 2025\\
**\- \[Introducing Auto Mode for the Parallel Task API\](/content/blog/task-api-auto-mode/index.html)**](/content/blog/task-api-auto-mode/index.html)

[Tags:](/content/blog/task-api-auto-mode/index.html) [Product](/content/blog?tag=product/index.html)

Author: By Parallel

[Jul 31, 2025\\
**\- \[A state-of-the-art search API purpose-built for agents\](/content/blog/search-api-benchmark/index.html)**](/content/blog/search-api-benchmark/index.html)

[Tags:](/content/blog/search-api-benchmark/index.html) [Benchmarks](/content/blog?tag=benchmarks/index.html)

Author: By Parallel

[Jul 31, 2025\\
**\- \[Parallel Search MCP Server in Devin\](/content/blog/parallel-search-mcp-in-devin/index.html)**](/content/blog/parallel-search-mcp-in-devin/index.html)

[Tags:](/content/blog/parallel-search-mcp-in-devin/index.html) [Product](/content/blog?tag=product/index.html)

Author: By Parallel

[Jul 28, 2025\\
**\- \[Introducing Tool Calling via MCP Servers\](/content/blog/mcp-tool-calling/index.html)**](/content/blog/mcp-tool-calling/index.html)

[Tags:](/content/blog/mcp-tool-calling/index.html) [Product](/content/blog?tag=product/index.html)

Author: By Parallel

[Jul 14, 2025\\
**\- \[Introducing the Parallel Search MCP Server \](/content/blog/search-mcp-server/index.html)**](/content/blog/search-mcp-server/index.html)

[Tags:](/content/blog/search-mcp-server/index.html) [Product](/content/blog?tag=product/index.html)

Author: By Parallel

[Jul 8, 2025\\
**\- \[Introducing Source Policy\](/content/blog/source-policy/index.html)**](/content/blog/source-policy/index.html)

[Tags:](/content/blog/source-policy/index.html) [Product](/content/blog?tag=product/index.html)

Author: By Parallel

[Jul 2, 2025\\
**\- \[The Parallel Task Group API\](/content/blog/task-group-api/index.html)**](/content/blog/task-group-api/index.html)

[Tags:](/content/blog/task-group-api/index.html) [Product](/content/blog?tag=product/index.html)

Author: By Parallel

[Jun 17, 2025\\
**\- \[State of the Art Deep Research APIs\](/content/blog/deep-research-browsecomp/index.html)**](/content/blog/deep-research-browsecomp/index.html)

[Tags:](/content/blog/deep-research-browsecomp/index.html) [Benchmarks](/content/blog?tag=benchmarks/index.html)

Author: By Parallel

[Jun 10, 2025\\
**\- \[Parallel Search API is now available in alpha\](/content/blog/search-api-alpha/index.html)**](/content/blog/search-api-alpha/index.html)

[Tags:](/content/blog/search-api-alpha/index.html) [Product](/content/blog?tag=product/index.html)

Author: By Parallel

[May 29, 2025\\
**\- \[Introducing the Parallel Chat API \](/content/blog/chat-api/index.html)**](/content/blog/chat-api/index.html)

[Tags:](/content/blog/chat-api/index.html) [Product](/content/blog?tag=product/index.html)

Author: By Parallel

[May 16, 2025\\
**\- \[Introducing Basis with Calibrated Confidences \](/content/blog/introducing-basis-with-calibrated-confidences/index.html)**](/content/blog/introducing-basis-with-calibrated-confidences/index.html)

[Tags:](/content/blog/introducing-basis-with-calibrated-confidences/index.html) [Product](/content/blog?tag=product/index.html)

Author: By Parallel

[Apr 24, 2025\\
**\- \[Introducing the Parallel Task API\](/content/blog/parallel-task-api/index.html)**](/content/blog/parallel-task-api/index.html)

[Tags:](/content/blog/parallel-task-api/index.html) [Product](/content/blog?tag=product/index.html),[Benchmarks](/content/blog?tag=benchmarks/index.html)

Author: By Parallel