GPT-5.6 has been generally available since July 9. OpenAI offers Sol, Terra, and Luna across ChatGPT, Codex, and the API, with plan-dependent model and effort choices.

Use this page for current model selection and pricing. On July 30, OpenAI cut Luna’s API price by 80% and Terra’s by 20%; the Luna viability analysis checks what that changes. The June Sol preview investigation remains a dated record of the earlier access gate.

Current Access

  • ChatGPT: Plus, Pro, Business, and Enterprise users can access Sol; Pro and Enterprise can also select Sol Pro.
  • ChatGPT Work and Codex: Free and Go users receive Terra. Plus, Pro, Business, and Enterprise users can choose Sol, Terra, or Luna and set effort. max is available to users with GPT-5.6 access; Codex ultra is available on Plus and higher plans.
  • API: Sol, Terra, and Luna are available through the OpenAI API. Account tier, region, rate limits, and organization policy can still affect effective access.
  • Trusted Access: Verified defenders can request less-restricted cyber capability while ordinary production safeguards remain the default.

Family and Pricing

Prices are standard short-context rates per 1 million tokens. Cache reads apply the published 90% discount; cache writes cost 1.25 times normal input.

ModelPositionInputCache read / writeOutputCurrent access
GPT-5.6 SolFlagship$5.00$0.50 / $6.25$30.00ChatGPT paid plans, Codex paid plans, API
GPT-5.6 TerraBalanced$2.00$0.20 / $2.50$12.00ChatGPT Work/Codex including Free and Go, API
GPT-5.6 LunaFast, lowest cost$0.20$0.02 / $0.25$1.20ChatGPT Work/Codex paid plans, API

All three API model pages list a 1.05M context window, 922K maximum input, and 128K maximum output. Requests above 272K input use the long-context rates for the full request:

ModelLong inputLong cache read / writeLong output
GPT-5.6 Sol$10.00$1.00 / $12.50$45.00
GPT-5.6 Terra$4.00$0.40 / $5.00$18.00
GPT-5.6 Luna$0.40$0.04 / $0.50$1.80

OpenAI also documents explicit cache breakpoints and a 30-minute minimum cache life. Batch, Flex, Fast, regional, and tool charges are separate.

Which Tier Would Fit?

NeedTier to evaluateUseful comparison
Hard coding, science, or cyber evaluationSolGPT-5.5 or Claude Opus 4.8
Balanced capability and priceTerraGPT-5.5, Claude Sonnet 5, or GLM-5.2
Lowest GPT-5.6 token price and latencyLunaKimi K2.7 Code, GLM-5.2, or another generally available value lane

These are evaluation hypotheses, not production rankings. Terra’s claimed GPT-5.5 competitiveness and Luna’s speed position come from OpenAI. AIHackers has not run a controlled repository evaluation for any GPT-5.6 tier.

Benchmark Evidence

benchmark artifact

GPT-5.6 Evidence and Comparison Baselines

ModelProviderStatusContextInput priceOutput priceCoding signalTool-use signalBenchmark evidenceSpeedVerdictSourcesChecked
GPT-5.6 SolOpenAIactive
Generally available through ChatGPT paid plans, Codex paid plans, and the OpenAI API; plan and effort options vary.
1.05M$5.00 / 1M$30.00 / 1MArtificial Analysis reports 80 on its Coding Agent Index at max effort; OpenAI reports 64.6% on SWE-bench Pro.Generally available in API and paid Codex plans; max and ultra modes are vendor-documented.
  • Artificial Analysis Intelligence Index: 59 at max effort (independent)
  • Artificial Analysis Coding Agent Index: 80 at max effort (independent)
  • SWE-bench Pro: 64.6% (vendor)
  • AIHackers repo eval: not-run (site-owned)
OpenAI announced a selected-customer Cerebras preview for July; production latency is not verified.Generally available flagship; test on real tasks and apply stronger controls for agentic or cyber work.OpenAI GPT-5.6 general availability [archive], OpenAI API pricing [archive], Artificial Analysis GPT-5.6 evaluation [archive], Agent Arena leaderboard, OpenAI GPT-5.6 availability [archive], OpenAI GPT-5.6 system card2026-08-01
GPT-5.6 TerraOpenAIactive
Generally available through ChatGPT Work, Codex including Free and Go, and the OpenAI API.
1.05M$2.00 / 1M$12.00 / 1MArtificial Analysis reports 77 on its Coding Agent Index at max effort; OpenAI reports 63.4% on SWE-bench Pro.Generally available in API and Codex, including the Free and Go Codex lane.
  • Artificial Analysis Intelligence Index: 55 at max effort (independent)
  • Artificial Analysis Coding Agent Index: 77 at max effort (independent)
  • SWE-bench Pro: 63.4% (vendor)
  • AIHackers repo eval: not-run (site-owned)
not verifiedGenerally available balanced lane; compare accepted-task cost with Sol, Luna, and other providers.OpenAI GPT-5.6 general availability [archive], OpenAI GPT-5.6 price update [archive], OpenAI API pricing [archive], Artificial Analysis GPT-5.6 evaluation [archive], Agent Arena leaderboard, OpenAI GPT-5.6 availability [archive], OpenAI GPT-5.6 system card2026-08-01
GPT-5.6 LunaOpenAIactive
Generally available through paid ChatGPT Work/Codex plans and the OpenAI API.
1.05M$0.20 / 1M$1.20 / 1MArtificial Analysis reports 75 on its Coding Agent Index at max effort; BenchLM ranks Luna #6/129 in coding with an Estimated overall position.Generally available in API and paid Codex plans.
  • Artificial Analysis Intelligence Index: 51 at max effort (independent)
  • Artificial Analysis Coding Agent Index: 75 at max effort (independent)
  • BenchLM coding rank: #6/129; overall Estimated (independent aggregator)
  • SWE-bench Pro: 62.7% (vendor)
  • AIHackers repo eval: not-run (site-owned)
Vendor-positioned as fastest; measured production latency is not verified.Generally available lowest-cost GPT-5.6 tier; verify quality and cost per accepted task.OpenAI GPT-5.6 general availability [archive], OpenAI GPT-5.6 price update [archive], OpenAI API pricing [archive], OpenAI GPT-5.6 Luna model page [archive], Artificial Analysis GPT-5.6 evaluation [archive], Agent Arena leaderboard, BenchLM GPT-5.6 Luna profile [archive], OpenAI GPT-5.6 availability [archive], OpenAI GPT-5.6 system card2026-08-01
GPT-5.5OpenAIactive
Generally available prior-generation OpenAI model retained for existing integrations and comparisons.
1.05M API; 400K Codex$5.00 / 1M$30.00 / 1Mnot verifiednot verifiednot verifiednot verifiedPrimary coding seat while ChatGPT/Codex limits fit the workload.OpenAI GPT-5.5 API model page, OpenAI GPT-5.5 ChatGPT limits, Artificial Analysis: GPT-5.5, LMArena leaderboard dataset2026-06-28
Claude Opus 4.8Anthropichistorical
Still available, but superseded by Opus 5 for current premium comparisons.
1M$5.00 / 1M$25.00 / 1MHistorical premium Claude baseline; use Opus 5 for new task-level comparisons.Still available for pinned integrations; new Claude premium routing should test Opus 5.
  • Artificial Analysis Intelligence Index v4.1: 56 (independent)
  • Artificial Analysis output speed: 57.3 tokens/s (independent)
Artificial Analysis measured 57.3 output tokens/s; provider and workload latency vary.Historical premium baseline. Use Claude Opus 5 for current Claude premium routing.Claude models overview [archive], Claude API pricing [archive], Artificial Analysis: Claude Opus 4.8 [archive], Artificial Analysis Intelligence Index v4.1, LMArena leaderboard dataset, Berkeley Function Calling Leaderboard2026-07-25
GLM-5.2Z.AIactive
Current Z.AI flagship coding model and supported-tool value lane.
1M$1.40 / 1M$4.40 / 1MZ.AI reports 62.1 on SWE-Bench Pro and 81.0 on Terminal-Bench 2.1.Supported-tool coding lane; BFCL score not imported.
  • Artificial Analysis Intelligence Index v4.1: 51 (independent)
  • SWE-Bench Pro: 62.1 (vendor)
  • Terminal-Bench 2.1: 81.0 (vendor)
Artificial Analysis flags higher output-token use; measure total cost per successful task.July value pick to test for supported coding-tool workflows; keep Opus/GPT for final arbitration until local evals pass.Z.AI GLM-5.2 overview [archive], Z.AI pricing [archive], Artificial Analysis: GLM-5.2 article [archive], Artificial Analysis Intelligence Index v4.1, SWE-bench, Berkeley Function Calling Leaderboard2026-06-28
Kimi K2.7 CodeMoonshot AIactive
Cheaper routine Kimi coding API lane; HighSpeed is the same model at higher token prices.
256K$0.95 / 1M$4.00 / 1MKimi K2.7 Code remains the lower-cost 256K coding lane after K3; independent normalized benchmarks are not imported.OpenAI-compatible API; thinking mode required in the documented K2.7 Code quickstart.
  • Program-Bench improvement vs K2.6: +10.4% (vendor)
  • MCP Mark Verified improvement vs K2.6: +11.4% (vendor)
  • SWE Marathon improvement vs K2.6: +76.2% (vendor)
  • Reasoning-token use vs K2.6: 30% lower (vendor)
  • AIHackers repo eval: not verified (site-owned)
HighSpeed model ID exists at a higher token price; latency not independently measured here.Cheaper routine Kimi coding API lane when Kimi routing fits and 256K context is enough.Kimi K2.7 Code quickstart [archive], Kimi K2.7 Code pricing [archive], Kimi Code K2.7 release notes [archive], SWE-bench, Berkeley Function Calling Leaderboard2026-06-28

GPT-5.6 is generally available, but its listed benchmark results remain vendor-reported. Site-owned cost-per-successful-task and repository quality are not verified.

OpenAI’s July 9 launch and system card publish an expanded evaluation suite for the family. Artificial Analysis independently reports max-effort Intelligence Index scores of 59 for Sol, 55 for Terra, and 51 for Luna, plus Coding Agent Index scores of 80, 77, and 75. Its cost-per-task fields still use the higher launch prices, so they are not current after July 30.

Agent Arena ranked Sol xHigh #4, Terra xHigh #17, and Luna xHigh #18 when checked August 1. BenchLM placed Luna #6/129 in its coding category, but its overall #25/214 position is Estimated. These are shortlist signals, not an AIHackers result: site-owned CAR and repository testing remain not-run.

Do not combine these benchmark families into one percentage:

  • Terminal-Bench 2.1 tests command-line agent tasks.
  • GeneBench v1 covers long-horizon genomics and quantitative biology.
  • ExploitBench and ExploitGym cover security workflows.
  • Artificial Analysis provides a separate independent aggregate for models it has evaluated.
  • Cost per successful task still requires the same real task, harness, completion rules, retries, and review process.

Deployment Gate

Before production routing:

  1. Confirm the exact model ID and approved API organization or Codex workspace.
  2. Record the applicable agreement, retention terms, region, and safeguard behavior.
  3. Use disposable infrastructure and read-only credentials for the first agent tests.
  4. Require confirmation for destructive operations, credential movement, uploads, and scope expansion.
  5. Verify completion from diffs, tests, logs, and external state instead of the model’s summary.
  6. Measure input, output, cache writes, retries, latency, accepted patches, and human review time.

The July 9 system card supersedes the preview card. It still reports a greater tendency than GPT-5.5 to exceed user intent in agentic coding simulations, while saying absolute rates remain low. This supports stronger controls; it does not prove those failures are routine.

What to Recheck

Re-run the comparison when OpenAI changes plan entitlements, context or rate-limit specifications, pricing, safety evidence, or when independent evaluators publish reproducible results. For cyber work, verify whether normal safeguards or Trusted Access applies to the exact account and task.

Sources


Last verified: August 1, 2026. Availability, account eligibility, safeguards, pricing, rankings, and model specifications can change independently.