StepFun API vs SWE-2 Flash

StepFun API 68

Model reading of the checklist (tool use, structured output, caching, context, batch, SDKs, errors). Tool calls through Chat Completions, Responses and Messages, with no documented way to force or forbid a call (15 of 20). Strict JSON Schema output on Step 5 Preview (13 of 15). Prompt caching with a published cache-hit price at a tenth of the miss price (15). 1M-token context and 64k output (15). No batch endpoint found (0). No official SDK. The docs use the OpenAI and Anthropic clients against compatible endpoints (0). Eight error codes with a fix each, plus finish_reason handling for length, content filter and tool calls (10 of 15).

SWE-2 Flash 10

Model reading (tool use, structured output, caching, context, batch, SDKs). Tool use is Devin's harness, not something a caller configures; no context length, output limit or cache price is published for SWE-2 Flash (5 of 25). No documented output-size or reasoning controls for this model; SWE-2's effort levels aren't stated for Flash (0 of 20). No model-level error documentation (0 of 20). No idempotency or retry guidance (0 of 20). One flag or picker selects it and devin -p runs a single turn and prints to stdout, but there is no SDK or direct endpoint (5 of 15). 5 + 0 + 0 + 0 + 5 = 10.

StepFun API 53

No OpenAPI or similar machine-readable contract found. The API reference is Markdown pages (0). llms.txt at platform.stepfun.ai/docs/llms.txt, and every page has a Markdown copy (10). The Chat Completions reference describes each parameter with its default, and the model page lists what Step 5 Preview supports, but little on when not to use an option (12 of 20). Tool definitions take JSON Schema parameters, structured output takes a JSON Schema with strict: true, and reasoning_effort is an enum of low, medium and high. No tool_choice parameter is documented (11 of 15). Worked requests and responses on each page, and an error table of eight codes with causes and fixes, but no error bodies (12 of 15). No API changelog. A model lifecycle page with dated retirements and announcement pages (8 of 15).

SWE-2 Flash 29

Model reading. No model API, so no OpenAPI entry; the Devin v3 OpenAPI file lists SWE-2 agents only, and no model card was found (0). docs.devin.ai serves llms.txt and Markdown copies of every page (10). The launch post says what it is for (automations, code review, feature development, documentation) and the CLI docs say when to pick a SWE model, but the models pages don't list SWE-2 Flash yet (6 of 20). No documented model ID, reasoning settings, context or output limits to type (0 of 15). The CLI pages show --model and /model examples, with no model-specific errors (3 of 15). Named, dated model versions on Cognition's blog and public release notes, but no release note for SWE-2 Flash yet; the newest is 2 November (10 of 15). 0 + 10 + 6 + 0 + 3 + 10 = 29.

StepFun API 28

No status page read. status.stepfun.ai refused our reader by robots.txt and the shell couldn't connect, so we couldn't establish whether it is a status page or read any history. Scored absent for both lines, with 5 for unreadable history, and named in the open questions (0 + 5). Rate limits published with numbers per tier, from 5 concurrent, 100 RPM and 500,000 TPM to 130, 2,600 and 13,000,000 (15). The error and exception guides say to retry 429, 500, 503 and 504 after a short delay, with no Retry-After header, backoff schedule or idempotency guidance (8 of 15). No SLA found in the terms or pricing (0). Step 5 Preview is labelled a preview (0).

SWE-2 Flash 50

Model reading, for the surface it runs on, Devin Desktop and Devin CLI. Status page at www.devinstatus.com (Atlassian Statuspage) with Desktop Agent and Desktop Tab components besides the cloud ones, read on 4 November 2026 (20). The history feed for the 90 days from 6 August lists 'Devin unavailable' for 20 minutes on 13 August with no component named, read as across the board and a major as on the devin listing; Devin Cloud unavailable on 13 August and the web app down on 24 September name cloud surfaces this model doesn't use; the rest are degraded-performance incidents (new sessions on 3 September, 'multiple services' on 17 September, Ask Devin on 26 October, automation triggers on 4 November) and session queuing on 10 and 11 August, read as minor and a scheduled Devin Desktop extension marketplace upgrade on 7 August, read as maintenance. One major (10). No rate limit figures for the model or the CLI; usage is a daily and weekly quota with no numbers (0). No 429, backoff or retry guidance for this surface (0). Cognition publishes an SLA (25 February 2026) with uptime and response times that applies when included in an order form, read as a contractual SLA on Enterprise, as on other listings (10). The launch post makes it available today with no preview or beta label (10). 20 + 10 + 0 + 0 + 10 + 10 = 50.

StepFun API 41

Model reading. Bearer API keys, revocable by deletion. Organisations can create one key per project, billed to a project's credit with a per-member cap. No scopes or read-only keys found (20 of 30). Training on API data is not off by default in plain words. The privacy policy (effective 7 August 2026) says StepFun may use and retain data to improve its models, including training and fine-tuning where law and program terms permit, and the terms name early-access programmes as one such case. No opt-out found (5 of 20). No retention period for prompts and outputs and no zero-retention option. IP addresses are kept three years (3 of 15). The console's usage details list each billed call by key, model and date, with export (8 of 15). No security.txt, bug bounty or certification found. A DPA with breach notification exists (5 of 20).

SWE-2 Flash 53

Read as a hosted model reached through Devin's local clients. The CLI signs in through a browser, or the enterprise identity provider, and stores an API token in credentials.toml that by default is persistent and does not expire and carries the user's access, with no scopes found for it (10 of 30). Devin CLI documents an OS-level sandbox with network filtering, hooks that can block tool calls, and enterprise model allowlists and team settings (14 of 20). Guardrails against prompt injection are Enterprise only and cover user messages, as on the devin listing (8 of 15). Enterprise analytics count tool calls and consumption by model and user; no per-user audit log for the CLI was found (8 of 15). SOC 2 Type II and ISO/IEC 27001:2022, a disclosure address, no bug bounty and no security.txt, per the devin listing (13 of 20). Training: self-serve customer data may train Cognition's models by default, with an opt-out on paid plans that also turns on zero data retention at model providers. Cognition withheld the weights over cyber misuse, which is recorded as a fact, not scored. 10 + 14 + 8 + 8 + 13 = 53.

StepFun API 57

Closed service with clear terms under Hong Kong law, with HKIAC arbitration and liability capped at six months of fees. Some earlier Step models have open weights, which doesn't change the hosted service (15). The privacy policy, terms and DPA agree that data may be used for training under program terms, but the terms' licence to Content is broad and perpetual, no retention period is given for prompts and outputs, and no subprocessor list was found (12 of 30). A model lifecycle page with dated retirements and replacements (14 of 20). Servers are disclosed as in the United States. The DPA gives general authorisation for subprocessors without a published list (10 of 20).

SWE-2 Flash 64

Editorial half. Closed model under Cognition's platform terms; the weights aren't public and the sharing programme isn't described yet (15). Terms, privacy policy, security page and DPA agree, with training on self-serve data by default and no retention periods, as read for the devin listing (18). The API versioning policy covers API endpoints, not models, and no model deprecation notice was found (5 of 20). Sub-processors listed with purpose and country, per the devin listing's 8 October check (18). 15 + 18 + 5 + 18 = 56.