Claude Opus 4.8 is now a historical integration and benchmark reference. It remains available, but Claude Opus 5 replaced it as the practical premium Claude baseline on July 24, 2026.

Do not use this page to justify new default routing to Opus 4.8. Use it when maintaining a pinned integration or reading dated comparisons; test Sonnet 5 first for routine work and Opus 5 for consequential premium passes.

Quick Facts

SpecClaude Opus 4.8
ProviderAnthropic
Model familyClaude Opus
Context1M-class in current Claude docs
API pricing$5 input / $25 output per 1M tokens
Batch pricing$2.50 input / $12.50 output per 1M tokens
Independent signalArtificial Analysis Intelligence Index 56; 57.3 output tokens/s at max effort in the current v4.1 snapshot
Best useHistorical comparison and pinned integrations
CaveatSuperseded by Opus 5 for current premium routing

Where Opus Fits

Start with cheaper or flatter-rate lanes when the work is routine:

WorkloadStart withEscalate to Opus 4.8 when…
Routine editsSonnet 5, GLM-5.2, Kimi K2.7 CodeThe first model misses behavior or introduces risky churn
Repo auditGLM-5.2 or MiniMax M3 when 1M context helpsThe answer needs premium reasoning or a final risk review
Code reviewGLM-5.2, Sonnet, or GPT-5.5 depending on tool fitYou need a high-confidence second pass
Migration planningGPT-5.5, GLM-5.2, or SonnetArchitecture tradeoffs are unclear or expensive to reverse
Compliance-sensitive analysisApproved provider path firstClaude terms and retention meet the workload’s policy needs

GLM-5.2 Comparison

Opus 4.8 was the premium Claude reference point for this comparison. The historical API cost gap was large:

ModelInput / output per 1MCost read
GLM-5.2$1.40 / $4.4072% lower input and 82.4% lower output than Opus 4.8
Claude Opus 4.8$5.00 / $25.00Premium review and arbitration lane

This is why GLM-5.2 is the July value pick to test. The test is not “does GLM win a chart?” The test is “does GLM solve your routine repo tasks well enough that Opus can move to final review?”

Subscription comparisons need a separate label. A $18 GLM Coding Lite plan is about 91% lower than a $200 Claude Max plan, but that is subscription math, not API pricing. Keep the lanes separate.

What Not To Claim

  • Do not call GLM-5.2 an Opus replacement without AIHackers-owned repo evals.
  • Do not treat “90% cheaper” as API pricing.
  • Do not route Fable as a normal premium alternative merely because access is restored; verify safeguards, retention, and cost. Mythos remains trusted-access only.
  • Do not use Opus 4.5 benchmark rows as the current Claude baseline.

Current Benchmark Evidence

benchmark artifact

Opus 4.8 Against Active and Restricted Lanes

ModelProviderStatusContextInput priceOutput priceCoding signalTool-use signalBenchmark evidenceSpeedVerdictSourcesChecked
Claude Sonnet 5Anthropicactive
Generally available across Claude plans, Claude Code, the Claude API, GitHub Copilot, and supported AWS paths.
1M$2.00 / 1M$10.00 / 1MAnthropic reports substantial coding and agentic gains over Sonnet 4.6; independent normalized results are pending.Available in Claude Code and the Claude API; adaptive thinking is on by default.
  • Cross-model benchmark evidence: vendor-reported; updated chart and system card preferred (vendor)
  • Artificial Analysis task cost: $1.53 per Intelligence Index task at max (independent)
  • AIHackers repo eval: not verified (site-owned)
No site-owned normalized latency result is verified.First Claude cost/performance test before Opus 5; escalate only when the premium pass changes the accepted result.Anthropic Claude Sonnet 5 launch [archive], Claude Sonnet 5 migration guide [archive], GitHub Copilot Claude Sonnet 5 launch [archive], Claude Sonnet 5 on AWS [archive], Artificial Analysis: Claude Opus 5 [archive]2026-07-25
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
Claude Fable 5Anthropicactive
Generally available; temporary subscription allowances ended July 7 and current subscription use is through usage credits.
1M$10.00 / 1M$50.00 / 1MAnthropic reports frontier launch results; independent reproducible ranking is pending.Guarded-domain requests can refuse or fall back; verify account behavior before routing.
  • Artificial Analysis Intelligence Index: 60 at max (independent)
  • AA-Briefcase: 1574 Elo / $22.30 per task (independent)
  • AIHackers repo eval: not verified (site-owned)
Task latency varies; compare complete-task runtime before escalation.High-cost guarded escalation only; use Opus 5 as the practical Claude premium baseline.Claude models overview [archive], Claude API pricing [archive], Anthropic Fable 5 and Mythos 5 [archive], Anthropic Fable/Mythos access statement [archive], Anthropic Fable 5 redeployment [archive], Artificial Analysis: Claude Opus 5 [archive], Artificial Analysis: Claude Opus 5 on AA-Briefcase [archive]2026-07-25
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
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

Artificial Analysis values are independent signals; Z.AI and Kimi coding claims remain vendor-labeled. AIHackers-owned cost-per-successful-task results are not yet verified.

Eval Pairing

Use the same task through both lanes:

TestGLM-5.2 pass signalOpus 4.8 role
Real bugCorrect patch, minimal churn, tests run or clearly scopedArbitration if GLM misses behavior
RefactorPreserves local conventions across 2-4 filesArchitecture review
Long-context auditAccurate repo map without invented filesHigh-confidence risk pass
ReviewConcrete file-grounded findingsFinal review before merge

If GLM-5.2 passes routine tasks, keep Opus for the cases where the premium pass visibly changes the outcome.

Sources


Last verified: July 25, 2026. Opus 4.8 remains available but is no longer the current premium comparison baseline.