GitHub

Executive Summary

What it is: GitHub Copilot is an AI-powered coding agent integrated into VS Code, Visual Studio, JetBrains, Xcode, Neovim, Eclipse, Zed, Raycast, the GitHub Copilot CLI, github.com, and the GitHub Copilot desktop app. It spans inline completions, chat, agent mode, automated code review, a cloud agent, and remote sessions controllable from GitHub Mobile. Individual plans run from $0 (Free) to $100/mo (Max), with Business at $19/user/mo and Enterprise at $39/user/mo. As of July 31, Copilot exposes 30+ models across seven providers: Anthropic, OpenAI, Google, Microsoft, GitHub (Raptor), xAI, and Moonshot AI. Source: https://github.com/features/copilot/plans

What to watch out for: July was a model-explosion month: seven new models shipped in 30 days, including Claude Opus 5 (Jul 24), the full GPT-5.6 family (Jul 9), Grok 4.5 (Jul 28), Gemini 3.6 Flash (Jul 21), and Kimi K2.7 Code (Jul 1). The pricing structure did not change, but the model lineup shifted dramatically. GPT-5.6 Luna is now the cheapest model in the lineup at $0.20/$1.20 per MTok, undercutting GPT-5.4 nano ($0.20/$1.25). Meanwhile, Gemini 2.5 Pro and Gemini 3 Flash are being deprecated on July 31. The bigger picture: the UK Competition and Markets Authority is probing Microsoft over Copilot-linked price hikes (reported Jul 29), and enterprise governance is tightening with a new "default model enablement" policy that turns all future GA models on by default for Business/Enterprise starting August 26. Source: https://www.bloomberg.com/news/articles/2026-07-29/microsoft-faces-uk-probe-over-copilot-linked-price-hikes , https://github.blog/changelog/2026-07-29-default-model-enablement-for-copilot-business-and-enterprise

Bottom line: Copilot's value proposition in July is unchanged from June (prepaid API access, not flat-rate), but the model menu is far richer. The addition of GPT-5.6 Luna gives budget-conscious users a genuinely cheap agentic option, while Opus 5 and GPT-5.6 Sol compete for the frontier. The real risk for enterprises is governance drift: the August 26 default-enablement policy will auto-expose every new GA model unless admins explicitly opt out. For individuals, the credit-burn problem from June persists with no published remedy from GitHub.

Key Terms

Latest Changes

Changes since the 2026-06 report.

Verification of June watch-items

  • Business/Enterprise promo credits ($30/$70): still pending. The pricing page still shows Business at $19/user/mo with $30 promo credits and Enterprise at $39/user/mo with $70 promo credits, with no change to the stated August 2026 expiration. The post-promo included amount (base only, or a new flex figure) remains undisclosed. Source: GitHub – Plans
  • Flex allotment variability: confirmed stable (no change). Flex amounts are unchanged from June: Pro +$5, Pro+ +$31, Max +$100. No reduction or increase was applied in July. The flex allotment is still explicitly variable. Source: GitHub – Plans
  • New models in July: confirmed. Seven new models shipped (see below). The June prediction of "Opus 5" materialized on July 24, and GPT-5.6 Sol/Terra/Luna landed July 9, exceeding the watch-item scope.
  • Copilot CLI updates: confirmed and expanded. June's real code intelligence (language server integration, Jun 26) was followed by a new C++ language server config skill (Jul 1), task-based auto model selection routing (Jul 1), per-session AI credit limits (Jul 1), and removal of the PAT requirement when using Copilot CLI in GitHub Actions (Jul 2). Sources: GitHub – 2026 07 01 New C Language Server Config Skill For Copilot Cli , GitHub – 2026 07 01 Copilot Cli Auto Model Selection Routes Based On Task , GitHub – 2026 07 02 Copilot Cli No Longer Needs A Personal Access Token In Github Actions

New July items

Plans

Individual Plans

Plan Price Total AI Credits / mo Base / Flex Key Inclusions
Free $0 Limited chat/agent (auto mode only); 2,000 completions/mo; 50 chat requests/mo n/a Haiku 4.5, GPT-5 mini, Copilot CLI, agent mode, MCP servers
Pro $10/mo $15 $10 base + $5 flex Cloud agent, code review, unlimited completions/next-edit, access to third-party agents (Claude Code, Codex), model selection (30+ models)
Pro+ $39/mo $70 $39 base + $31 flex Premium models (Opus 4.5-5, GPT-5.5, GPT-5.6 Sol), audit logs, GitHub Spark (preview), delegate to third-party coding agents (preview), 4x+ included usage vs Pro
Max $100/mo $200 $100 base + $100 flex Priority access to new models and features, 2.9x+ included usage vs Pro+, all Pro+ features

Source: GitHub – Plans , GitHub – Github Copilot Individual Plans Introducing Flex Allotments In Pro And Pro And A New Max Plan

Business / Enterprise Plans

Plan Price Included AI Credits / mo (through Aug 2026) Key Inclusions
Business $19/user/mo $30 promo ($19 base + promo) All models, SSO, policy management, IP indemnity, code review, cloud agent, pooled credits, admin budgets, default model enablement (Aug 26)
Enterprise $39/user/mo $70 promo ($39 base + promo) Everything in Business + codebase indexing, fine-tuned models, Spaces API, audit logs, per-user credit budgets, managed-settings.json, OpenTelemetry export, MDM deployment

Source: GitHub – Github Copilot Is Moving To Usage Based Billing , GitHub – 2026 07 29 Default Model Enablement For Copilot Business And Enterprise

Notes: Code completions and next-edit suggestions remain unlimited and do not consume credits on any paid plan. Pooled usage (Business/Enterprise) shares included credits org-wide with admin spend controls. Additional AI Credits can be purchased on Pro, Pro+, and Max (not Free); purchase is blocked for subscriptions made via GitHub Mobile on iOS/Android. The Free plan includes 2,000 completions and 50 chat requests (including Copilot Edits) per month. Source: GitHub – Models And Pricing

API Pricing

Copilot does not sell standalone API access; it meters subscription usage at each model's published per-million-token rate, where the dollar amount equals the number of AI Credits consumed (1 credit = $0.01). All figures are per 1M tokens, from GitHub's models-and-pricing reference.

OpenAI

Model Input Cached input Output Notes
GPT-5.6 Luna (<=200K) $0.20 $0.02 $1.20 New cheapest model; long context >200K: $0.40/$0.04/$1.80
GPT-5.4 nano $0.20 $0.02 $1.25 Former cheapest
GPT-5 mini $0.25 $0.025 $2.00 Lightweight
Raptor mini (GitHub fine-tuned) $0.25 $0.025 $2.00
GPT-5.4 mini $0.75 $0.075 $4.50
GPT-5.6 Terra (<=272K) $2.00 $0.20 $12.00 Balanced default; long context: $4.00/$0.40/$18.00
GPT-5.3-Codex $1.75 $0.175 $14.00 LTS base model for Business/Enterprise
GPT-5.4 (<=272K) $2.50 $0.25 $15.00 Long context >272K: $5.00/$0.50/$22.50
GPT-5.5 (<=272K) $5.00 $0.50 $30.00 Long context: $10.00/$1.00/$45.00
GPT-5.6 Sol (<=272K) $5.00 $0.50 $30.00 Highest reasoning; long context: $10.00/$1.00/$45.00

Anthropic

Model Input Cached input Cache write Output Notes
Claude Haiku 4.5 $1.00 $0.10 $1.25 $5.00
Claude Sonnet 5 (promo to Aug 31) $2.00 $0.20 $2.50 $10.00 Rises to $3.00/$0.30/$3.75/$15.00 after promo
Claude Sonnet 4 / 4.5 / 4.6 $3.00 $0.30 $3.75 $15.00
Claude Opus 4.5 / 4.6 / 4.7 / 4.8 / 5 $5.00 $0.50 $6.25 $25.00 Opus 5 added Jul 24, same price
Claude Opus 4.8 (fast mode) (preview) $10.00 $1.00 $12.50 $50.00 2x standard Opus 4.8
Claude Fable 5 $10.00 $1.00 $12.50 $50.00 Listed but currently unavailable (export-control suspension)

Google

Model Input Cached input Output Notes
Gemini 3 Flash (preview) $0.50 $0.05 $3.00 Deprecated Jul 31; use 3.5 Flash
Gemini 3.5 Flash $1.50 $0.15 $9.00 Lightweight
Gemini 3.6 Flash $1.50 $0.15 $7.50 New; cheaper output than 3.5 Flash
Gemini 2.5 Pro $1.25 $0.125 $10.00 Deprecated Jul 31; use 3.1 Pro
Gemini 3.1 Pro (preview, <=200K) $2.00 $0.20 $12.00 Long context >200K: $4.00/$0.40/$18.00

xAI

Model Input Cached input Output Notes
Grok 4.5 (<=200K) $2.00 $0.50 $6.00 500K context window; long context >200K: $4.00/$1.00/$12.00

Moonshot AI

Model Input Cached input Output Notes
Kimi K2.7 Code $0.95 $0.19 $4.00 First Moonshot model in Copilot

Microsoft

Model Input Cached input Output Notes
MAI-Code-1-Flash $0.75 $0.075 $4.50 First Microsoft homegrown model

Source: GitHub – Models And Pricing

Other pricing notes:

  • Copilot code review is an exception to transparent per-token cost: the review model is auto-selected and not disclosed, so per-review cost varies and cannot be estimated from the table. Reviews also consume undisclosed GitHub Actions minutes. Source: GitHub – Models And Pricing
  • Annual Pro/Pro+ subscribers on legacy request-based billing keep model multipliers (a different scheme), not the per-token rates above. Source: GitHub – Model Multipliers For Annual Plans
  • GPT-5.2 and GPT-5.2-Codex still appear in the model availability list on the plans page but do not appear in the per-token pricing table, meaning their cost per interaction is not publicly documented.

Model Performance / Benchmarks

GitHub did not publish a new quantitative benchmark evaluation in July. The June 25 harness evaluation (comparing the GitHub Copilot agentic harness against model-vendor harnesses across SWE-bench Verified, SWE-bench Pro, SkillsBench, Win-Hill, and TerminalBench 2.0) remains the most recent. The July 27 blog post "The harness is all you need (mostly)" reinforced the harness-centric thesis qualitatively: that Copilot's multi-model orchestration (automatically dispatching a small "Explore" subagent for reads and a larger "General Purpose" subagent for complex actions) is the productivity differentiator, not any single model.

The blog also recommends GPT-5.6 Terra or Claude Sonnet on medium reasoning as the default for most work, citing prompt caching benefits from model consistency within a session. The Rubber Duck review workflow (cross-model review) is positioned as a quality differentiator that single-model tools cannot replicate. Source: GitHub – The Harness Is All You Need Mostly

External model-level signals (not Copilot-specific, but relevant to model selection within Copilot):

  • GPT-5.6 Sol generated significant HN interest (1,561 points, 1,113 comments) on its July 9 launch, with users reporting strong reasoning performance on complex tasks. Source: News – Item
  • A production migration report found GPT-5.6 delivered 2.2x faster responses at 27% lower cost versus the prior model in an agentic workload. Source: Ploy – Migrating A Production Ai Agent To Gpt 5 6

GitHub does not publish a standalone Copilot SWE-Bench score; performance depends on the selected model and harness configuration.

Latest News

Seven New Models Land in July (Jul 1-28)

July saw the largest single-month model expansion in Copilot history. The GPT-5.6 family (Sol/Terra/Luna, Jul 9) added a three-tier OpenAI lineup with Luna as the new cheapest model at $0.20/$1.20. Claude Opus 5 (Jul 24) matched Opus 4.x pricing at $5.00/$25.00. Grok 4.5 (Jul 28) brought xAI into Copilot with a 500K context window. Gemini 3.6 Flash (Jul 21) undercut its predecessor's output cost. Kimi K2.7 Code (Jul 1) added Moonshot AI as the seventh provider. Sources: GitHub – 2026 07 09 Openais Gpt 5 6 Sol Terra And Luna Are Now Available In Github Copilot , GitHub – 2026 07 24 Claude Opus 5 Is Now Available In Github Copilot , GitHub – 2026 07 28 Grok 4 5 Is Now Available In Github Copilot

Default Model Enablement Policy for Business/Enterprise (July 29)

GitHub announced that GA models will be turned on by default for Copilot Business and Enterprise plans starting August 26, replacing the per-model manual approval workflow. A 28-day review window (Jul 29 to Aug 26) lets admins set a global opt-out. Open-weight models (DeepSeek, Kimi K2.7) and models without GitHub data-retention coverage (Fable 5) are excluded from auto-enablement. Source: GitHub – 2026 07 29 Default Model Enablement For Copilot Business And Enterprise

Gemini 2.5 Pro and Gemini 3 Flash Deprecated (July 31)

Both models are removed from all Copilot experiences on July 31, 2026. Alternatives: Gemini 3.1 Pro and Gemini 3.5 Flash. This is the first batch model deprecation since the AI Credits launch. Source: GitHub – 2026 07 02 Upcoming Deprecation Of Gemini 2 5 Pro And Gemini 3 Flash

Copilot Code Review Gets Agent Skills and MCP (July 29)

Code review reached GA support for agent skills and MCP servers, allowing reviews to incorporate external tools and custom analysis beyond static diff inspection. Source: GitHub – 2026 07 29 Copilot Code Review Agent Skills And Mcp Now Generally Available

Copilot App, Vision, and Browser Tools Reach GA (July 1-7)

The standalone GitHub Copilot desktop app became available to all users (Jul 7). Copilot vision (image processing in chat) and browser tools (web interaction in VS Code) both reached GA on July 1. Source: GitHub – 2026 07 07 Github Copilot App Available To All , GitHub – 2026 07 01 Copilot Vision Is Generally Available

Enterprise Governance Push: Managed Settings, MDM, OpenTelemetry (July 1-27)

GitHub shipped a wave of enterprise management features: managed-settings.json GA (Jul 1), MDM deployment for VS Code/CLI (Jul 8), enterprise-managed OpenTelemetry export (Jul 8), managed settings for the Copilot app and cloud agent (Jul 27), a dedicated Copilot app access policy (Jul 27), and granular AI credit budgeting at the cost-center level (Jul 2, 7, 20). Sources: GitHub – 2026 07 01 Enterprise Managed Settings Json Is Generally Available , GitHub – 2026 07 08 Enterprise Managed Opentelemetry Export For Vs Code And Cli

UK Probe Into Microsoft Over Copilot Price Hikes (July 29)

Bloomberg reported that the UK competition authority is probing Microsoft over Copilot-linked price increases, the first regulatory scrutiny of the AI Credits billing model. The article is paywalled. Source: Bloomberg – Microsoft Faces Uk Probe Over Copilot Linked Price Hikes

GitHub Models Fully Retired (July 30)

The standalone GitHub Models product (an API playground distinct from Copilot) was fully retired on July 30, 2026. Source: GitHub – 2026 07 01 Github Models Is Being Fully Retired On July 30 2026

Community Signals

UK regulatory scrutiny of Copilot pricing

The Bloomberg report (Jul 29) that the UK competition authority is investigating Microsoft over Copilot-linked price hikes was submitted to HN and generated discussion. This is the first known regulatory action tied to the AI Credits billing transition.

  • sbulaev submitted the story "Microsoft Faces UK Probe over Copilot-Linked Price Hikes" (8 points). News – Item

Credit burn remains the dominant complaint, now with concrete numbers

The June narrative of "blew through a month of credits in days" continued into July, with users providing increasingly specific cost breakdowns.

  • curao_d_espanto calculated the economics in detail: "Copilot Pro+ having 7000 'AI credits' monthly is insanely low considering that a single gpt5.4 request would use 250 credits (28 request per month), and 5.5/opus using 500 (and even more depending on the output size!), which is even 14 request per month." News – Item
  • hathym, commenting on the Kimi K2.7 Code GA announcement: "Same here, I've removed my credit card from Copilot and won't be renewing." News – Item

Users actively switching to alternatives

A consistent theme across July HN comments is developers migrating from Copilot to direct API access (OpenRouter, OpenCode) or competing agents (Claude Code, Cursor) to avoid the credit markup.

  • Computer0, on the MAI-Code-1-Flash launch thread: "I went to VSC specifically to avoid the pricing I started experiencing on Cursor. After this change I have no reason to stick with GH Copilot, I'd rather keep buying OR credits." News – Item
  • pkaye offered a counterpoint, noting that cheaper models like Haiku 4.5 are viable for interactive coding if the developer understands the codebase: "Because the Haiku model is quite cheap but doesn't screw up too often I used it for interactive coding... there were some months when I used GitHub copilot where I had excess credits available at the end of the month I frantically try to use up." News – Item

GPT-5.6 generates massive excitement on HN

The July 9 launch of GPT-5.6 (and its Sol variant) was one of the highest-engaged HN stories of the month, signaling strong demand for the model within Copilot.

  • The main GPT-5.6 thread reached 1,561 points and 1,113 comments. News – Item
  • A production migration report by Ploy found "Migrating a production AI agent to GPT-5.6: 2.2x faster, 27% cheaper" (258 points). News – Item

Security: AI worms target Microsoft Copilot (not GitHub Copilot)

Late July saw multiple disclosures of AI worm vulnerabilities in Microsoft 365 Copilot (Copilot for Word), where malicious documents could self-propagate through Copilot sessions. These affect Microsoft's M365 Copilot product, not GitHub Copilot, but are relevant to any organization deploying Copilot broadly.

  • Canopy9560: "Document-borne AI worms can self-propagate through Copilot for Word" (380 points, 294 comments). News – Item
  • wunderwuzzi23: "Copirate 365: Plundering in the Depths of Microsoft Copilot (CVE-2026-24299)" (a DEF CON talk). News – Item

Failed sources

  • The Bloomberg article on the UK probe (Bloomberg – Microsoft Faces Uk Probe Over Copilot Linked Price Hikes) returned HTTP 403 (paywalled). The headline and submission details were obtained from the HN thread. Full details of the probe's scope and findings could not be verified.
  • The Reddit JSON endpoints (old.reddit.com/r/GithubCopilot/...json) returned HTTP 403 for multiple Copilot pricing threads, so July Reddit sentiment could not be fetched programmatically. HN coverage was used instead.

Enterprise Readiness

Feature Available? Details
SSO (SAML) Yes Business plan and above. Individual plans (including Pro+/Max) explicitly exclude SAML SSO. Source: GitHub – Plans
SSO (OIDC) Partial OIDC via GitHub's SAML integration; not a standalone option.
SCIM Yes Enterprise plan. Source: GitHub – Identity And Access Management
Audit logs Yes Pro+ (individual) and above; Enterprise has full org/enterprise audit logs. Source: GitHub – Plans
IP indemnity Yes Business and Enterprise, when content filtering is enabled, under Microsoft's Copilot Copyright Commitment. Not included on any individual plan. Source: GitHub – Plans , Blogs – Copilot Copyright Commitment Ai Legal Concerns
Data residency Partial GitHub Enterprise Cloud supports EU data residency. No air-gapped option.
HIPAA Yes Enterprise plan with BAA. Source: GitHub – Enterprise
Air-gapped / on-prem No GitHub Enterprise Server can run on-prem, but Copilot agent features require cloud connectivity.
SLA Yes GitHub Enterprise Cloud offers a 99.9% uptime SLA. Source: GitHub – Enterprise
Admin controls (RBAC) Yes Policy management, model restrictions, org-level settings, pooled-credit budgets, per-user AI credit budgets, managed-settings.json (GA Jul 1), MDM deployment (Jul 8), enterprise-managed OpenTelemetry (Jul 8), default model enablement (effective Aug 26). Sources: GitHub – Plans , GitHub – 2026 07 01 Enterprise Managed Settings Json Is Generally Available , GitHub – 2026 07 29 Default Model Enablement For Copilot Business And Enterprise

Transparency Gaps

Gap Details Severity
Business/Enterprise post-promo amount The $30 (Business) / $70 (Enterprise) promo credits expire after August 2026. The ongoing included amount after the promo (base only, or a new flex) is not stated. With the promo ending imminently, this is the highest-risk unknown for enterprise budgeting. High
Default model enablement governance risk Starting August 26, all future GA models turn on by default for Business/Enterprise. While there is an opt-out, organizations that do not act within the 28-day window will silently gain new models, potentially including high-cost frontier models, without admin review. The cost impact of auto-enabled models is not quantified. High
Copilot code review cost The review model is auto-selected and not disclosed, so per-review token cost cannot be estimated from the pricing table. Reviews also consume undisclosed GitHub Actions minutes. High
Real-world credit burn rate No published estimate of how many credits a typical agentic session (e.g. 2 hours with Opus 5) consumes; community reports of draining a month's credits in 1-2 days remain uncorroborated by GitHub. The Jul 20 feature letting users see credits per billing cycle helps visibility but does not provide forward-looking estimates. High
UK probe scope The UK competition authority probe into Copilot-linked price hikes was reported but is paywalled. The specific allegations, scope, and potential outcomes are not publicly documented. Medium
GPT-5.2 / GPT-5.2-Codex pricing These models appear in the model availability list on the plans page but do not appear in the per-token pricing table, so their cost per interaction is undocumented. Medium
Flex allotment instability Flex credits ($5/$31/$100 for Pro/Pro+/Max) are explicitly variable and "may change over time," so total included usage is not guaranteed beyond the base. July saw no change, but this remains an unresolved budget risk. Medium
Free tier auto-mode model set Free/Student are locked to auto model selection, but which models auto routing actually selects (and their cost tiers) is not documented. Medium
Opus 5 cyber-content safeguards Opus 5 includes enhanced safeguards that "may block some cyber-related or security-adjacent requests." The specific triggers and scope of blocking are not documented, which could affect security engineering workflows. Medium
Model count vs. plan gating complexity With 30+ models across 7 providers and different availability tiers per plan (e.g., GPT-5.6 Sol and Opus 5 are Pro+ only, while Luna and Terra are Pro-accessible), the cost implications of model selection are increasingly hard to predict without consulting the full pricing table. Low
MAI-Code-1-Flash scope Microsoft's first homegrown model is limited to Business/Enterprise and has no published benchmarks or context-window spec on the pricing page. Low