Gemini 2.5 Flash is Google’s stable hybrid reasoning profile for well-rounded multimodal work and agentic tool use. It offers thinking and non-thinking responses for large-scale multimodal tasks that need different levels of reasoning. Text, images, audio and video contribute to a text answer. Document input is also described in Google’s model-specific cloud profile. A useful result can be a document comparison, an interpreted chart, a transcript, code or a structured record.
The edition’s reasoning control uses thinkingBudget. Dynamic thinking is the default; a setting of -1 lets the model adapt its reasoning to task complexity. An explicit budget can range from zero to 24,576 tokens, and zero turns thinking off. Choose a budget appropriate to the actual problem rather than assigning the same effort to a simple categorization and a multi-condition analysis. The guide treats the budget as guidance, and the native card notes that adherence may not be exact.
For a document or data task, define the question, required fields and evidence that should accompany the result. For a requirements comparison, compare two specifications and return the changed requirement, its source passage and the effect on a stated workflow. Structured output is useful when each finding needs the same record shape. If the source gives several possible values or leaves a field open, preserve that uncertainty instead of forcing a single plausible answer. Review returned values against the supplied pages or data.
For visual or audio material, identify the relevant object, label, event or recording passage. Ask for the kind of text needed: visible values from a chart, a description of an event, a transcript or a summary. These artifacts are different, and the instruction should say which one to produce. The shared native card includes Flash Image and Flash Audio as separately named capabilities; their generated image and speech outputs are not part of this stable text profile.
Google’s exact cloud profile documents PDF and plain-text inputs with interface-specific limits: 50 MB per PDF imported through the API and up to 3,000 pages per file, with a 7 MB plain-text file limit. That profile also describes approximately 45 minutes of video with audio or one hour without audio, and approximately 8.4 hours of audio within the million-token input allowance. These are the stated cloud-interface constraints, separate from the general model token window. Prepare the relevant section and source format rather than assuming every file type or upload path has the same limit.
The exact Gemini API table lists function calling, code execution, file search, search grounding, Maps grounding, URL context and structured outputs. Define each function’s role and parameter shape and continue from the real returned result. A model can request an operation, while the connected application performs it. Preserve the necessary thought-signature context when continuing a reasoning task in the stateless interface. A structured answer or a function suggestion does not independently establish that an external operation succeeded.
The native model card gives 1M context and 64K text output, while the exact API table specifies 1,048,576 input tokens and 65,536 output tokens. A response cap includes thinking tokens, so an overly small cap can leave the final answer truncated or empty. Ask for an artifact sized for the output allowance and use the reasoning budget to guide effort. Caching, batch, Flex and Priority processing are documented for this producer profile. Retain item identifiers when preparing several independent records.
The stable edition remains distinct from Flash Preview variants and later Gemini 3 models. Its card lists possible hallucinations, limitations in causal understanding and complex logical or counterfactual reasoning, and a January 2025 knowledge cutoff. Provide fresh evidence when a task depends on newer information. Check calculations, extracted values, transcripts and proposed code against the actual source or execution result. Hybrid thinking is useful when it supports that concrete result, rather than becoming a substitute for reviewing it.
", "aboutText": "Gemini 2.5 Flash is Google’s stable hybrid reasoning profile for well-rounded multimodal work and agentic tool use. It offers thinking and non-thinking responses for large-scale multimodal tasks that need different levels of reasoning. Text, images, audio and video contribute to a text answer. Document input is also described in Google’s model-specific cloud profile. A useful result can be a document comparison, an interpreted chart, a transcript, code or a structured record.\n\nThe edition’s reasoning control uses thinkingBudget. Dynamic thinking is the default; a setting of -1 lets the model adapt its reasoning to task complexity. An explicit budget can range from zero to 24,576 tokens, and zero turns thinking off. Choose a budget appropriate to the actual problem rather than assigning the same effort to a simple categorization and a multi-condition analysis. The guide treats the budget as guidance, and the native card notes that adherence may not be exact.\n\nFor a document or data task, define the question, required fields and evidence that should accompany the result. For a requirements comparison, compare two specifications and return the changed requirement, its source passage and the effect on a stated workflow. Structured output is useful when each finding needs the same record shape. If the source gives several possible values or leaves a field open, preserve that uncertainty instead of forcing a single plausible answer. Review returned values against the supplied pages or data.\n\nFor visual or audio material, identify the relevant object, label, event or recording passage. Ask for the kind of text needed: visible values from a chart, a description of an event, a transcript or a summary. These artifacts are different, and the instruction should say which one to produce. The shared native card includes Flash Image and Flash Audio as separately named capabilities; their generated image and speech outputs are not part of this stable text profile.\n\nGoogle’s exact cloud profile documents PDF and plain-text inputs with interface-specific limits: 50 MB per PDF imported through the API and up to 3,000 pages per file, with a 7 MB plain-text file limit. That profile also describes approximately 45 minutes of video with audio or one hour without audio, and approximately 8.4 hours of audio within the million-token input allowance. These are the stated cloud-interface constraints, separate from the general model token window. Prepare the relevant section and source format rather than assuming every file type or upload path has the same limit.\n\nThe exact Gemini API table lists function calling, code execution, file search, search grounding, Maps grounding, URL context and structured outputs. Define each function’s role and parameter shape and continue from the real returned result. A model can request an operation, while the connected application performs it. Preserve the necessary thought-signature context when continuing a reasoning task in the stateless interface. A structured answer or a function suggestion does not independently establish that an external operation succeeded.\n\nThe native model card gives 1M context and 64K text output, while the exact API table specifies 1,048,576 input tokens and 65,536 output tokens. A response cap includes thinking tokens, so an overly small cap can leave the final answer truncated or empty. Ask for an artifact sized for the output allowance and use the reasoning budget to guide effort. Caching, batch, Flex and Priority processing are documented for this producer profile. Retain item identifiers when preparing several independent records.\n\nThe stable edition remains distinct from Flash Preview variants and later Gemini 3 models. Its card lists possible hallucinations, limitations in causal understanding and complex logical or counterfactual reasoning, and a January 2025 knowledge cutoff. Provide fresh evidence when a task depends on newer information. Check calculations, extracted values, transcripts and proposed code against the actual source or execution result. Hybrid thinking is useful when it supports that concrete result, rather than becoming a substitute for reviewing it.", "editorialOverlay": { "copy_fields": null, "copy_sha256": null, "copy_source": null, "html_sha256": "fdb8f4821ed15c52fe9db9112db5081c4db3e1b571fc7b28e995fd1d23d6f99e", "module_sha256": "ee9dc1c50b1e6a8fb7fa062030fcb4646b9aaaeb08be513b723a5a92a7290f06", "raw_model_unchanged": true, "schema": "ailoot-native-model-editorial-v1" }, "locale": "en", "methods": [ { "html": "google/gemini-2.5-flash:batchGemini 2.5 Flash is Google’s stable hybrid reasoning model for multimodal analysis and agent tasks. Its dynamic thinking budget can be set from 0 to 24,576 tokens, including an off mode. Text, images, audio, video and documents become text or structured results, with a documented 1,048,576-token input limit.
This description covers the model’s capabilities. Available in Ailoot inputs, outputs and options are listed for the selected method below.
Gemini 2.5 Flash is Google’s stable hybrid reasoning profile for well-rounded multimodal work and agentic tool use. It offers thinking and non-thinking responses for large-scale multimodal tasks that need different levels of reasoning. Text, images, audio and video contribute to a text answer. Document input is also described in Google’s model-specific cloud profile. A useful result can be a document comparison, an interpreted chart, a transcript, code or a structured record.
The edition’s reasoning control uses thinkingBudget. Dynamic thinking is the default; a setting of -1 lets the model adapt its reasoning to task complexity. An explicit budget can range from zero to 24,576 tokens, and zero turns thinking off. Choose a budget appropriate to the actual problem rather than assigning the same effort to a simple categorization and a multi-condition analysis. The guide treats the budget as guidance, and the native card notes that adherence may not be exact.
For a document or data task, define the question, required fields and evidence that should accompany the result. For a requirements comparison, compare two specifications and return the changed requirement, its source passage and the effect on a stated workflow. Structured output is useful when each finding needs the same record shape. If the source gives several possible values or leaves a field open, preserve that uncertainty instead of forcing a single plausible answer. Review returned values against the supplied pages or data.
For visual or audio material, identify the relevant object, label, event or recording passage. Ask for the kind of text needed: visible values from a chart, a description of an event, a transcript or a summary. These artifacts are different, and the instruction should say which one to produce. The shared native card includes Flash Image and Flash Audio as separately named capabilities; their generated image and speech outputs are not part of this stable text profile.
Google’s exact cloud profile documents PDF and plain-text inputs with interface-specific limits: 50 MB per PDF imported through the API and up to 3,000 pages per file, with a 7 MB plain-text file limit. That profile also describes approximately 45 minutes of video with audio or one hour without audio, and approximately 8.4 hours of audio within the million-token input allowance. These are the stated cloud-interface constraints, separate from the general model token window. Prepare the relevant section and source format rather than assuming every file type or upload path has the same limit.
The exact Gemini API table lists function calling, code execution, file search, search grounding, Maps grounding, URL context and structured outputs. Define each function’s role and parameter shape and continue from the real returned result. A model can request an operation, while the connected application performs it. Preserve the necessary thought-signature context when continuing a reasoning task in the stateless interface. A structured answer or a function suggestion does not independently establish that an external operation succeeded.
The native model card gives 1M context and 64K text output, while the exact API table specifies 1,048,576 input tokens and 65,536 output tokens. A response cap includes thinking tokens, so an overly small cap can leave the final answer truncated or empty. Ask for an artifact sized for the output allowance and use the reasoning budget to guide effort. Caching, batch, Flex and Priority processing are documented for this producer profile. Retain item identifiers when preparing several independent records.
The stable edition remains distinct from Flash Preview variants and later Gemini 3 models. Its card lists possible hallucinations, limitations in causal understanding and complex logical or counterfactual reasoning, and a January 2025 knowledge cutoff. Provide fresh evidence when a task depends on newer information. Check calculations, extracted values, transcripts and proposed code against the actual source or execution result. Hybrid thinking is useful when it supports that concrete result, rather than becoming a substitute for reviewing it.
POST /v1/batchestexttext| Field | Type | Requirement | Values & limits | Default |
|---|---|---|---|---|
Fieldendpoint | Typestring | RequirementRequired | Values & limits /v1/chat/completions | Default— |
Fieldmodel | Typestring | RequirementRequired | Values & limits google/gemini-2.5-flash:batch | Default— |
Fieldrequests | Typebatch_item[] | RequirementRequired | Values & limits Minimum items: 1 | Default— |
Fieldrequests[].custom_id | Typestring | RequirementRequired | Values & limits | Default— |
Fieldrequests[].body.messages | Typemessage[] | RequirementRequired | Values & limits | Default— |
Fieldcompletion_window | Typestring | RequirementOptional | Values & limits 24h | Default— |
Fieldprovider | Typeobject|null | RequirementOptional | Values & limits | Default— |
| Parameter | Contract |
|---|---|
Parametercontent_type | Contractapplication/json |
Parameterrequired | Contractendpoint, model, requests |
Parametermaterial_types | Contracttext |
Parameterrepresentation | Contractinline requests[]; each item has custom_id and body |
Parameterrequest_endpoint_profile | Contract/v1/chat/completions |
Parameterfile_material | ContractCaller may read actual UTF-8 JSONL locally into requests[]. No upload purpose, input_file_id or file reference is used. Missing/empty customer material stops preparation before POST. |
Parameterserialization_order | Contractendpoint, model, provider (if supplied), completion_window (if supplied), requests last |
Parameteroptional_controls | Contractprovider.only, provider.allow_fallbacks, completion_window |
Parameterschema_scope | ContractText Chat Completions bodies in the documented inline Batch envelope. |
Lifecycle: asynchronous
GET /v1/batches/{operation_id} · actual original create response.id; same user/key/group
Tariff references do not add request options. Supported parameters are listed for the selected method.
Match each charge to the conditions shown. Use one rate for each set of alternative configurations.
| Configuration | Rate schedule | Context tokens, more than | Context tokens, up to and including | Charge |
|---|---|---|---|---|
| ConfigurationBase token rates | Rate scheduleBase rate | Context tokens, more than— | Context tokens, up to and includingNo upper bound | Charge Input · Input tokens $0.0003 / 1K tokens Output · Output tokens $0.0025 / 1K tokens Input · Cache read $0.00006 / 1K tokens Input · Cache write — / 1K tokens Input · Cache write, 1 hour — / 1K tokens |
| ConfigurationRequest-length tier 1 | Rate schedule— | Context tokens, more than0 | Context tokens, up to and includingNo upper bound | Charge Input · Input tokens $0.0003 / 1K tokens Output · Output tokens $0.0025 / 1K tokens Input · Cache read $0.00006 / 1K tokens Input · Cache write — / 1K tokens Input · Cache write, 1 hour — / 1K tokens |
Ailoot reserves an estimated amount before a request. After the result, it settles the final charge from measured usage. Streaming settlement uses the final usage record. How reservation and settlement work →
set -eu\n: "${AILOOT_API_KEY:?Set your Ailoot API key}"\npython3 - <<'PY_REQUEST'\nimport json\nfrom pathlib import Path\nbody=json.loads('{"endpoint":"/v1/chat/completions","model":"google/gemini-2.5-flash:batch","requests":[{"custom_id":"req-0001","body":{"messages":[{"role":"user","content":"Hello!"}],"max_tokens":16}}]}')\nPath("batch-request.json").write_text(json.dumps(body,ensure_ascii=False),encoding="utf-8")\nPY_REQUEST\ncurl --fail-with-body --silent --show-error https://api.ailoot.ai/v1/batches \\\n -H "Authorization: Bearer $AILOOT_API_KEY" -H 'Content-Type: application/json' \\\n --data-binary @batch-request.json -D batch-response-headers.txt -o batch-create-response.json\nimport json, os\nfrom pathlib import Path\nfrom urllib.request import Request, urlopen\n\ndef prepare_requests(customer_path):\n path = Path(customer_path)\n if not path.is_file(): raise ValueError("Customer JSONL file is missing")\n items = [json.loads(line) for line in path.read_text(encoding="utf-8").splitlines() if line.strip()]\n if not items: raise ValueError("Customer JSONL contains no requests")\n ids = set()\n for item in items:\n if set(item) != {"custom_id", "body"}: raise ValueError("Each item needs custom_id and body")\n cid = item["custom_id"]\n if not isinstance(cid, str) or not cid.strip() or cid in ids: raise ValueError("custom_id must be nonempty and unique")\n ids.add(cid)\n if not isinstance(item["body"], dict) or not item["body"].get("messages"): raise ValueError("Supply actual text messages")\n return {"endpoint":"/v1/chat/completions", "model":"google/gemini-2.5-flash:batch", "requests":items}\n\nif __name__ == "__main__":\n material = os.environ.get("AILOOT_BATCH_REQUESTS_JSONL", "")\n if not material: raise SystemExit("Set AILOOT_BATCH_REQUESTS_JSONL to your actual customer file")\n body = prepare_requests(material)\n key = os.environ.get("AILOOT_API_KEY", "")\n if not key: raise SystemExit("Set AILOOT_API_KEY")\n request = Request("https://api.ailoot.ai/v1/batches", data=json.dumps(body, ensure_ascii=False).encode("utf-8"), headers={"Authorization":"Bearer " + key, "Content-Type":"application/json"}, method="POST")\n with urlopen(request) as response:\n raw = response.read()\n Path("batch-create-response.json").write_bytes(raw)\n print(raw.decode("utf-8"))\ninline results[].response.body JSON
GET /v1/batches/{operation_id}set -eu\n: "${AILOOT_API_KEY:?Set your Ailoot API key}"\nBATCH_OPERATION_ID="$(python3 - <<'PY_ID'\nimport json\nfrom pathlib import Path\nfrom urllib.parse import quote\nresponse = json.loads(Path("batch-create-response.json").read_text(encoding="utf-8"))\noperation = response.get("id")\nif not isinstance(operation, str) or not operation: raise SystemExit("No actual original operation ID; preserve the response and do not resubmit")\nprint(quote(operation, safe=""))\nPY_ID\n)"\ncurl --fail-with-body --silent --show-error "https://api.ailoot.ai/v1/batches/$BATCH_OPERATION_ID" \\\n -H "Authorization: Bearer $AILOOT_API_KEY" -o batch-status-response.json\nKeep every complete returned response, file and content part.
Keep every result, response/error body, choice and aggregate usage.
Preserve the original operation ID. 503 result_pending_check has X-Should-Retry:false; never resubmit an unknown original batch.
| Field | Type |
|---|---|
Fieldresults[].custom_id | Typestring |
Fieldresults[].response.body.choices[].message.content | Typestring|null |
Fieldresults[].error | Typeobject|null |
Fieldusage | Typeobject|null |
Ailoot Ultra google/gemini-2.5-flash:batch
Use the method paths shown above.